On pigeons and people: A preliminary look at the columban simulation project.

The author set up pigeon experiments that copied tricky human acts like "insight" and "self-awareness."

Birds worked for food in chambers with lights and keys. No fancy brain talk was used.

The paper is a progress report, not a full experiment. It lists several new simulations still running.

Pigeons acted like they suddenly "saw" the answer, even though only food schedules guided them.

The birds also showed signs of "knowing they know" without any inner mind needed.

In short, simple environmental rules made birds look smart.

Sanabria et al. (2009) extend the idea. Their pigeons played a Matching-Pennies game and copied the same small mistakes people make. A plain learning model explained both species.

Pisacreta (1982) followed one year later and showed birds can lock into long, rigid key-peck chains. That work gives a concrete example of the complex units the 1981 paper said could stand in for human "planning."

Szempruch et al. (1993) tried to build "transitive inference" in pigeons. They got fast learning but only spotty inference. This tempers the 1981 claim: some symbolic tasks need extra steps before the simulation works.

You can treat sudden client "insight" as the product of tiny, trackable reinforcement shifts. When a learner seems to "get it," check what just changed in timing, payoff, or cues instead of crediting a mental leap. Try slicing the task into smaller, clearer contingencies first.