The Sky Was Never a Clock
On what remains of an orrery when you remove the clockwork — and the stone that pays Le Verrier’s debt without being told whose debt it is.
The first orreries were honest about their metaphor. Brass planets on brass arms, geared to a crank: turn the handle and the heavens advance. The gearing was the theory — the solar system as the inside of a watch, God as the watchmaker, time as the shaft everything turns on. Three centuries later the brass is gone but the shaft remains: every simulation you have ever seen advances the planets by t, as if somewhere under the ecliptic a master clock were ticking and the planets were checking it.
They are not checking anything. Mercury does not know what time it is. It knows — in the only sense that matters to physics — where it is in its own closing, and the closings are what we actually count. Ask what a year is and you will not find a duration; you will find a ratio: while Earth closes once, Mercury closes 4.15 times, Jupiter a twelfth. Kepler wrote his third law with periods in it, but divide any two copies of it and the seconds cancel. What survives is bookkeeping: the square of the count against the cube of the orbit. Counts against counts. The ledger.
And the ledger, read closely, does something no metaphor could: its errors turn out to be inventory. Balance all eight rows and Jupiter sits a part in a thousand off — not because the law fails, but because Jupiter weighs a part in a thousand of the Sun, and the exact ledger says 1 + m∕M☉ where the schoolbook says 1. The audit weighs its auditors. In 1846 the imbalance in Uranus’s row was resolved the same way — with mass, not repeal — and the mass was Neptune, found within a degree of where the arithmetic said to look.
The hard test is the fall. It is easy to place planets on counted orbits; the question is whether the counting can earn one. So the new instrument drops a stone into the census raw: one line of calculus — radius as a function of its own turning, no clock as the parameter — integrated step by step, the second difference checked against the rule at every step, live, where you can watch it. The stone is given three numbers: the size of the orbit, its stretch, and the seam of the Sun. It is not given Mercury. It is not given 1859, or Le Verrier, or the 43 arcseconds a century that no Newtonian bookkeeping could pay and that Vulcan was invented — and never found — to cover. The ellipse closes. The perihelion walks forward. The machine reads its own lean: 42.98″ a century. The modern ephemerides, holding four decades of radar and the MESSENGER mission, say 42.98 ± 0.04.
Even the second law changes station. In the textbook, equal areas in equal times is a law to be proven. In a counted sky it cannot be a law, because there are no equal times to compare against — the swept area is how the count is read. Demote the clock and Kepler’s second law is promoted to a definition. Nothing about the sky changed; what changed is which line is load-bearing.
None of this is new physics, and the page flies that flag at full volume: every number the counting produces, the standard account produces — re-derivation is not discovery. But there is a difference between a picture of a watch and a census that balances, and the difference is the crank. The old orreries let you turn the heavens by hand and called the shaft time. This one hands you the ledger instead — and the ledger, unlike the shaft, was measured.