Differential autoshaping to common and distinctive elements of positive and negative discriminative stimuli.

Researchers taught pigeons to peck two keys. One key showed a picture with a special mark. The other key showed almost the same picture, but without the mark.

The birds got food only when they pecked the key with the special mark. Pecking the wrong key gave nothing. This setup is called autoshaping.

The team watched which parts of the pictures the birds pecked most. They wanted to know if the birds used the special mark or the shared background to choose.

The birds quickly learned to peck the key with the special mark. They pecked right near that mark during correct trials.

During wrong-key trials, they still pecked, but now they aimed at the shared background parts. The special mark guided correct choices, while the shared parts guided errors.

Henson et al. (1979) ran almost the same task. They saw the same pattern: birds use the odd feature first. Their birds needed extra training when the pictures were squashed together, but the end result matched.

Zentall et al. (1975) looked like they disagreed. They placed both pictures side-by-side at the same time and saw no clear mark-vs-background split. The difference is timing. A et al. showed pictures one after the other; R et al. showed them together. Same birds, same cues, different layout, different outcome.

Cohen et al. (1990) later added more features. They used five cues instead of two and still saw clean control by each separate part. The 1974 study opened the door; the 1990 study walked through with a heavier suitcase.

When you set up a discrimination task, spotlight the one feature that matters. Hide or minimize everything else. If the client must pick the red spoon, don’t also give red cups and red plates at first. Start with one clear difference, then slowly add shared parts once the key feature controls the response.