Journal · No. 003

120 songs tapped, three recognized

Christoph Svoboda · November 12, 2026 · 2 min

Stanford, 1990. Elizabeth Newton asked people to tap the rhythm of a well known song on a table. Their partner had to name it. 120 songs were tapped. 3 were recognized.

Before each round, the tappers estimated their listener's chances. They said one in two. It was one in forty.

The tappers said one in two. It was one in forty.

The tappers were not careless. While they tapped, they heard the melody in their head. The listener heard knocking. Only one of them knew the difference.

Anyone who has presented a strategy has been the tapper. Three rounds of slides, examples, a Q&A. Weeks later someone repeats it back, and what comes back is knocking.

That is why “is that clear?” settles nothing. The person explaining hears the song while explaining it. Their sense of how well it landed is not a weak measurement. It is wrong in one direction, and the direction is optimistic.

When did you last tap a song nobody else could hear?

In the glossary: Shared ReferenceInternal Coherence