01. The Fact Box — Every Number Had a First Person

What the entry does not tell you

An encyclopedia entry on the Earth gives its circumference as roughly 40,075 kilometres and moves on to the next paragraph. The number arrives finished, as though it had always been available, which is exactly the impression a reference work is built to give. But every number in every entry was once unknown, and for each one there was a first person who did not know it either and then did. That person almost always worked with equipment that looks absurd now: a stick in the ground, a lead ball on a wire, a horn antenna full of pigeons. The gap between the apparatus and the answer is the interesting part, and the fact box is where it is guaranteed not to appear. This book puts ten of those numbers back into the hands of the people who first held them.

Why the stories are never the tidy version

The tidy version says a clever person had an idea and measured the thing correctly. The real version is that Eratosthenes' accuracy depends on the length of a unit nobody can now pin down, so historians still argue about how right he was. It is that the metre was built on a survey containing an error its own author discovered, concealed, and carried to his death. It is that Cavendish never reported the constant everyone credits him with, and that the constant in question remains the worst-measured in physics. It is that Rømer published a light-time rather than a speed, and that the speed we quote was somebody else's arithmetic. Each chapter here keeps the complication rather than smoothing it, because the complication is what teaches the method.

Why a child was right to care

A child who reads encyclopedias for pleasure is usually accused of collecting trivia, which misreads what the pleasure is. The pleasure is the sense that the world has been found out — that somewhere a person stood in a field and worked out how far away the sun is. Learning how a number was made converts it from something to memorise into something to think with, and teaches the reader when to doubt one. That is the only durable defence against a confident figure with nothing behind it, a more common object than a well-made measurement.

Example: A cathedral seen from the pew is a finished thing; seen from the scaffolding it is a set of decisions, and only the second view lets you build one.

Verdict. A measured number is a claim with a method attached, and reading the method is what separates knowing a fact from possessing one.

Overturned if: these standard accounts survive contact with their primary sources unamended, with no concealed error, misattribution or disputed unit.

Sources. 2018 · Dmitry A. Shcheglov, "The so-called 'Itinerary Stade' and the Accuracy of Eratosthenes' Measurement of the Earth," Klio — De Gruyter · 2022 · CODATA, "Recommended Values of the Fundamental Physical Constants" — NIST