High-order concept formation in the pigeon.

Garcia (1974) taught pigeons to peck only when a photo showed a man-made object. Food followed pecks on man-made scenes. No food followed pecks on nature-only scenes.

After the birds learned the rule, new photos were shown. The test asked: do the birds still pick man-made objects they have never seen?

The pigeons kept choosing the man-made class with brand-new pictures. They had formed a high-order concept, not just memorized old slides.

The study gave early proof that simple reinforcement can create abstract visual categories in non-humans.

Vyazovska et al. (2016) extended the same lab set-up but asked birds to watch four visual dimensions at once. Both studies show pigeons can master complex go/no-go tasks, yet Vyazovska reveals attention trade-offs when cues multiply.

Qadri et al. (2026) pushed the photo idea into hospital work. Their birds learned to spot lung nodules in CT scans and generalized to unseen patients. The medical task is harder, yet the core procedure—peck for target, skip for non-target—mirrors Garcia (1974).

Kennedy et al. (1993) swapped man-made objects for moving objects. Pigeons still generalized, proving the effect is not tied to one stimulus type. Together these papers build a 50-year line showing reinforcement can craft many visual concepts.

You already use differential reinforcement to teach tacts, intraverbals, or safety responses. This pigeon work reminds you to test generalization with truly novel examples—new photos, new voices, new rooms—before claiming mastery. It also shows that clear yes/no consequences can build surprisingly abstract discriminations, a cue to keep your own teaching contingencies clean and consistent.