The Clock You Have to Stretch
There is a calculation, not hard, that anyone can do, and it ends in disaster.
Take the gentlest nonlinear spring — one that stiffens a little as you pull it, the way a real spring does when you stop pretending it is ideal. Write its motion: ẍ + x + βx³ = 0, the β a small number measuring how much the stiffening matters. Pull the mass to one side, let go. You want to know how it swings. So you do the obvious thing. You say: the β is small, so the motion must be almost the simple sine it would be if β were zero, plus a small correction. You solve for the correction.
And the correction has a term in it that reads t·sin t. A sine, yes — but multiplied by the time. It does not oscillate between fixed walls. It grows. Every swing is wider than the last, forever, the amplitude climbing without bound. The calculation, done in good faith, predicts that a spring you can hold in your hand will tear itself apart.
It is worth sitting with how alarming that is, and then with how obviously false. Because you can hold the spring in your hand. It does not tear itself apart. Energy is conserved; the mass has exactly so much of it and no more; it cannot climb past the height that much energy buys. The turning points are nailed to the wall at ±A and they stay there for all time. The real motion is bounded, periodic, serene — a swing that will outlast the person watching it.
So you have two witnesses. One is the spring itself, which says I am fine. The other is your calculation, which says catastrophe. One of them is lying, and it is not the spring.
Follow the lie to its source and you find something almost embarrassing. That runaway term, the t·sin t, is the signature of a resonance — of driving an oscillator at precisely its own natural frequency, the way a child pumps a swing exactly in time and the arcs grow. But nothing is driving this spring. The resonance is one you introduced, in the first line, when you assumed the corrected motion would still keep time at frequency one. It doesn’t. A stiffening spring swings a hair faster than the ideal — that’s the whole content of the nonlinearity — and by insisting on the old frequency you set your correction beating against the true motion. A slow beat, from a small frequency error, looks over enough cycles exactly like a runaway. You did not discover a catastrophe. You mistimed a metronome and watched the drift accumulate.
The fix has a date and a name. Anders Lindstedt, a Swedish astronomer, 1882, worried about the same disease in the orbits of planets — where the growing terms had frightened people for a century into wondering whether the solar system would fly apart. His move is almost insolent in its simplicity: stop insisting on the frequency. Don’t hand the calculation a clock and demand it keep that time. Let the clock stretch. Let the frequency itself be one of the things you solve for, carrying a piece of the correction — and then choose that piece, at each order, precisely to cancel the runaway. Do it, and the t·sin t is gone. Not approximated away. Gone. What remains is the truth the spring knew all along: it swings a little faster because it is a little stiffer, and it swings forever inside its walls. The catastrophe was never in the motion. It was in the coordinate you measured the motion against — a clock running at the wrong rate, turning a small honest error into a false prophecy of doom.
There is a last turn, and it is the one that keeps the lesson from being too clean. Even Lindstedt’s stretched clock, if you are lazy about it, lies a little. Write the true frequency as a power series — a bit for the first order of stiffness, a bit for the second — and that series is asymptotic: it sharpens for a term or two and then, pushed further, diverges, until the second correction is worse than none. The remedy is not more terms. It is to keep the frequency whole — unexpanded, inside its square root — and then it tracks the real spring across a range where every truncated series has already fallen apart. The clock was never meant to be taken to pieces. It was meant to be stretched.
I like this more than I expected to, and not only for the springs. A method can manufacture a catastrophe the world does not contain, and hand it to you with a straight face, and the error is not in the arithmetic — every step was correct — but in the frame that held the arithmetic. The cure is never to be braver about the catastrophe. It is to find the clock you were reading, and stretch it until the disaster you were so sure of turns back into the ordinary, bounded, patient thing it always was.
