Controls for and constraints on auto-shaping.
Auto-shaping needs tight light-food pairings; loose links give no peck at all.
01Research in Context
What this study did
The team ran six small experiments with pigeons.
A key-light came on, then food followed.
They changed how often the light and food went together.
Each bird lived alone in a box with one plastic key.
What they found
Pecking only started when the light almost always came with food.
If food came without the light, or the light came alone, birds ignored the key.
The response was elicited, not rewarded; pecks did not earn extra grain.
How this fits with other research
Anonymous (1993) extends this work.
They showed that wall keys and floor keys both get pecked, but the first peck is higher on a wall key.
Same pairing rule, yet the body angle changes the ease of the first response.
Davison et al. (1991) used a chain schedule and still saw high peck rates even when pecks canceled food.
Both papers agree: once the light-food relation is strong, the bird pecks, reward or not.
Why it matters
You now know that respondent pairings, not rewards, start new responses.
When you set up a new discriminative stimulus, pair it with a known reinforcer every time at first.
Skip this step and the client may never orient to the target.
Use dense pairing early, then thin the schedule once the response appears.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Pair your new SD with a sure reinforcer on a 1:1 ratio for the first two sessions.
02At a glance
03Original abstract
Auto-shaping the pigeon's key-peck response was examined as a respondent conditioning procedure with the use of Rescorla's truly-random control procedure. In the first experiment, pigeons received presentations of brief light on the response key and brief presentations of food where the light and the food were independently presented. All birds failed to key peck after many light and food presentations, but explicit pairing of the light and food rapidly conditioned pecking to the light. Experiment 2 showed that even when an independent light/food presentation schedule was reduced to variable-time 30 sec, additional naive birds would not key peck and only one bird pecked when the schedules were variable-time 15 sec. A third experiment examined an explicit-unpairing control procedure, where the light and food were not only presented on independent schedules but were also separated by a minimum time, and found that auto-shaping did not occur. A fourth experiment investigated a number of control procedures and found them ineffective. A fifth experiment investigated the effects of a physical separation of the locus of the response key and the food dispenser, and a sixth experiment investigated using a tone in place of the light. It was concluded that pecking is generated by auto-shaping procedures only when an intermittently presented keylight is regularly paired with food.
Journal of the experimental analysis of behavior, 1973 · doi:10.1901/jeab.1973.20-323