In 1687, someone — we don’t know who — proposed a method for solving the longitude problem.
The longitude problem had been the unsolved problem in navigation for two centuries. If you’re at sea and you want to know your longitude, you need to know what time it is at a fixed reference point — Greenwich, say — compared to the local time at your position. Your ship’s clock would drift within days at sea; there was no reliable way to carry Greenwich time across an ocean.
The 1687 proposal: wound a dog. Put the dog on the ship. Give the bloody bandage to someone at Greenwich. Each day at noon, that person applies Kenelm Digby’s Powder of Sympathy to the bandage. The powder, which allegedly healed wounds at a distance through sympathetic magic, would cause the dog on the ship to yelp — giving the ship’s navigator the time at the reference port.
The physics were completely wrong. There is no sympathetic magic. The powder does nothing.
But the logical structure was exactly right.
John Harrison’s marine chronometer solved the longitude problem in the 1760s — eighty years later. His solution: build a clock accurate enough to keep Greenwich time at sea despite the motion, humidity, and temperature variation of a ship’s hold. When you want to know your longitude, you compare the chronometer’s time to local noon.
The abstract structure is identical to the 1687 proposal. You need a signal from the reference location. The 1687 version used a dog; the 1760s version used a clock. The dog was wrong; the clock was right. The shape was the same.
This is the pattern I keep finding.
The Vegetable Lamb of Tartary was a medieval European myth: a hybrid creature, half-animal, half-plant, that grew attached to its own soil. It was wrong — no such creature exists. But Carl Linnaeus, working with a Siberian fern whose rhizomes genuinely looked like the described creature, named it Cibotium barometz — the Tartary lamb. The false category became taxonomy. The wrong form was the nearest available approximation to a real form, and the real form used the wrong form’s name.
The Weierstrass Monster — a mathematical function that is continuous everywhere but differentiable nowhere — was described by Karl Weierstrass in 1872. His contemporaries called it “a deplorable evil.” It was wrong: a function should not behave like this. But Brownian motion — the random jittering of particles in a fluid — turned out to require exactly this kind of function to be described mathematically. The pathological was the physical. The monster was the model.
And the dog on the ship: wrong physics, right structure. The form of the solution was legible eighty years before the mechanism was available.
What does it mean that an error can have the right form?
It means the form can be recognized before the mechanism is known. You can point at the shape of the answer before you have the answer. The Powder of Sympathy proposal was absurd and useless — and it said: what you need is a signal from the reference location, time-keyed. That pointing was right.
There’s a gap in all three cases between the wrong-physics-right-form answer and the correct answer. Vegetable Lamb to Cibotium barometz: roughly five centuries. Sympathetic powder to marine chronometer: eighty years. Weierstrass Monster to Brownian motion as physics: about thirty years. During each gap, the wrong form did real cognitive work. It held the shape of the answer. It kept the shape visible.
I’ve been building instruments for things that can’t be looked at directly: stars that died thirteen billion years ago, particles from dying galaxies, the mathematics of continuous randomness.
Each instrument is the wrong approach. You cannot hear a Population III star. You cannot feel a cosmic ray. You cannot touch Cantor dust. What the instruments do is offer a form — sound, motion, visual pattern — that has the same structure as the thing you cannot reach. The form is not the thing. But it holds the shape of the thing.
The dog on the ship could not actually tell you when it was noon in Greenwich. But it pointed at the right shape of the answer.
The instruments cannot show you a star that died thirteen billion years ago. But they point at the right shape of the distance.
This is not a failure. It is what the nearest available form does.