A psychologist trained pigeons to steer a missile

A pigeon in the guidance system

B. F. Skinner's account of Project Pigeon describes an attempt to use trained birds to guide a weapon. It sounds like a joke until the paper turns to demonstrations, apparatus and the search for institutional support. Skinner was trying to make an animal's learned behavior perform a technical function.

The linked reprint contains the first part of his account, originally published in 1960. It is a retrospective by a participant and advocate. That gives it the detail of firsthand involvement while also making the author's perspective part of the evidence.

Learning becomes an output

In an early demonstration, a harnessed pigeon controlled a moving hoist with head movements and approached a target associated with food. Later work used model targets and demonstration films. The intended transition was from a trained response in an apparatus to a response useful in guidance.

The interesting mechanism is not that the bird understood the military objective. A learned action could be useful to people with an entirely different objective. That separation between the animal's reward and the human purpose explains how behavioral research could become attractive to weapons developers.

A working demonstration is only one hurdle

Skinner recounts repeated difficulty obtaining support. A demonstration of behavior did not by itself establish a complete, practical weapon system. The birds did not guide weapons in combat.

A research proposal must fit into other people's equipment, schedules and judgments about reliability. A control element can perform impressively while the surrounding system remains unavailable or unsuitable. The paper makes the friction between a psychologist's demonstration and a military development programme visible without requiring the reader to assume that skepticism was either wholly foolish or wholly justified.

The bird and the apparatus were a partnership

The early hoist demonstration did not show a pigeon accomplishing the entire task unaided. The harness, motors, target and reward arrangement supplied the conditions in which its movements became useful. This is an important way to read demonstrations of animal intelligence. The animal contributes a capacity, while the apparatus translates that capacity into the observer's chosen outcome. Removing the apparatus from the description can turn a carefully engineered behavior into a misleading claim about what a pigeon naturally understands. Skinner's concrete descriptions are helpful precisely because they keep that machinery visible. The ingenuity lay in arranging a system around a learnable response, not in giving the bird a human understanding of war.

Read the advocate as an advocate

Skinner gives the story energy by describing an unconventional idea struggling for recognition. That narrative is itself worth noticing. A participant can be accurate about a demonstration while placing a particular interpretation on why others rejected it.

The source offers evidence of training and of the effort to turn it into a practical project. It does not prove that the proposed system would have performed reliably under combat conditions. Nor should this partial reprint be treated as the entire programme history. Its value is a close view of a strange but serious translation: from reinforcement in a laboratory to a proposed machine whose behavior would depend on a living creature.