Autoshaping with common and distinctive stimulus elements, compact and dispersed arrays.
Pack stimuli tight for quick first pecks, then spotlight one clear cue to lock in long-term control.
01Research in Context
What this study did
The team worked with pigeons in a small lab box. The birds saw two kinds of picture panels. One kind had many tiny dots packed tight. The other kind had the same dots spread out.
A bright star shape sat in the middle of every panel. If the bird pecked, food might drop. This is called autoshaping. The test ran for many days.
What they found
Dense, compact arrays taught the first pecks faster. Birds reached 100 pecks sooner with the tight pictures.
Later, the star shape alone guided most choices. But this switch only happened after long training. Early on, the star helped little.
How this fits with other research
Lazar (1977) showed that special feedback lights speed up new chain learning. E et al. now show the same idea works for upkeep after the skill is born.
Johnson et al. (1968) proved that past discrimination practice makes single cues stronger. The 1979 data match this: the star only took charge after many sessions.
Williams et al. (2002) pushed array work further. They found pigeons can tell same-from-different pictures with just 2-second looks. Together, these papers say array density matters for speed, but cue salience rules long-term control.
Why it matters
When you teach a new discrimination, pack the early stimuli tight to get responding fast. Once the learner pecks reliably, add or highlight one clear cue like the star. Keep that cue consistent across trials. This two-step plan shortens acquisition time and builds strong stimulus control that lasts.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Put the teaching cards close together for the first week, then space them out and add a bright border around the target card.
02At a glance
03Original abstract
Four groups of pigeons were trained with a standard autoshaping procedure in which a brief fixed-duration interval always followed by a grain delivery alternated with a longer variable-duration interval never associated with grain delivery. One of two stimuli was always presented during each interval. One of them contained three black dots and a black star on a green background; the other contained four black dots on a green background. The four elements of each stimulus were arranged in a more compact array for two groups and in a more dispersed array for the other two groups. Which of the two stimuli preceded grain delivery was counterbalanced within each pair of groups. The speed of occurrence of the first autoshaped peck was not affected by whether the stimulus containing the distinctive star element preceded grain delivery, but autoshaping was faster when the stimulus arrays were compact than when they were dispersed. During 560 response-independent training trials that followed the first autoshaped peck, this pattern reversed; both discriminative control over responding and the relative frequency of pecking the stimulus that preceded grain delivery were greater for the two groups where this stimulus contained the discriminative element than for the two groups where it contained only common elements. During subsequent testing with stimuli containing only a single element each, the distinctive feature was responded to proportionately more often by the two groups for which it had been an element of the stimulus preceding grain delivery than by the two groups for which it had been an element of the stimulus complex that never was associated with grain delivery. These data add further support to the hypothesis that the initial occurrence of autoshaped responding and its subsequent maintenance are not affected by the same variables. They also suggest that automaintenance is as sensitive as response-dependent training to the presence or absence of a distinctive stimulus element among several common elements and that this sensitivity appears to be independent of the specific method used for presenting the stimuli during automaintenance.
Journal of the experimental analysis of behavior, 1979 · doi:10.1901/jeab.1979.31-383