Pigeons' spatial memory: II. Acquisition of delayed matching of key location and transfer to new locations.

Researchers taught pigeons to peck a key that matched the location of a sample key. After a short delay, the birds had to pick the same spot again.

Once the birds could do this, the team moved the keys to brand-new places. They wanted to see if the pigeons would still pick the right spot without extra training.

Four out of five pigeons got it right the first time they saw the new spots. They used a general rule: 'match the place,' not just memorize each spot.

The birds kept choosing correctly even when the keys appeared in places they had never seen before.

Catania et al. (1972) seems to disagree. They gave pigeons only three of four key setups. When a new setup appeared, the birds failed. The difference is training range. Haemmerlie (1983) trained many spots from the start, so transfer was easy.

Sprague et al. (1984) later swapped color shapes for places. Their birds also matched new shapes right away. Together, the two papers show pigeons can learn broad rules for both space and shape.

Saunders et al. (1988) found pigeons did NOT show symmetry or transitivity after basic matching. This warns us that general matching does not always create full equivalence classes. You get transfer, but not always the full set of emergent relations.

When you teach a client to match, vary the item positions from day one. Use many spots on the table, screen, or wall. This builds a flexible 'match the place' rule instead of a narrow habit. Kids are not pigeons, but the principle holds: broad training now saves re-teaching later.