Why we invent patterns in random noise

The screen shows a large triangle and two smaller shapes moving around a box.

Viewer 1: “The big one is bullying them.”

Viewer 2: “The little triangle is brave.”

Viewer 3: “It trapped the other one.”

Thirty-five of thirty-six participants in the 1944 study described motives or personalities. Geometry acquired villains, courage and pursuit without changing a single angle.

Now shuffle the sequence. Reverse it. Remove the box. People produce another account. The stories differ, yet each feels discovered because every selected movement can be fitted after the fact.

This is a magnificent human ability. It lets us infer danger, intention and cause from incomplete scenes. It also lets three upward points become “momentum”, one difficult customer become “a segment”, and a coincidence become a strategy.

The useful evidence in the first interpretation is hardly that the triangle was aggressive. It is that aggression came easily to the observer. Another group might see protection, play or escape. The projection reveals the available story before it reveals the moving object.

An analyst watches the animation twice and writes “persistent pursuit”. On the reversed version, the same motion becomes “strategic retreat”. Both labels are internally coherent. Neither was present in the file. The exercise becomes less a test of shapes than an inventory of the explanations waiting in the viewer.

On the final replay, the large triangle pauses beside the opening. Somebody says it is waiting to attack. Somebody else says it is letting the smaller shape leave. The animation ends after a minute.

Behavioural principles

Behavioural ideas at play in this post

Short, plain-English explanations of the principles behind this post, with links to related books and examples in the archive.