Fourier analysis

Fourier's claim was audacious: take any function over an interval — even a jagged one with corners or jumps — and it equals an infinite sum of sines and cosines, each oscillating at a whole-number multiple of one base frequency, each present in a precise amount you can compute by an integral. A square wave, all flat tops and sharp edges, is the sum of a sine and its odd harmonics, fading in strength. Fourier found this modeling heat conduction — the temperature along a bar, governed by the heat equation — and laid it out in *Théorie analytique de la chaleur* (1822). Any shape, he was saying, is a chord: a stack of pure tones.

The idea was so sweeping it was resisted. Fourier submitted a first memoir to the Institut de France in 1807, and Lagrange — who had his own views on trigonometric series — blocked its publication; the treatise appeared only in 1822, after Lagrange had died. And Fourier was broadly right, but making his claim *precise* forced mathematics to grow up. Exactly when does the infinite series converge, and to what value at a jump? Answering that drove Dirichlet's convergence conditions, a sharper definition of what a "function" even is, and eventually the Riemann and Lebesgue integrals. A physics result about heat reshaped the foundations of analysis.

Fourier analysis is the frequency view of the world. It underlies signal processing, the compression of sound and images (MP3 and JPEG both discard frequencies the senses barely register), spectroscopy, the position–momentum duality of quantum mechanics, radio, and medical imaging like MRI. The fast Fourier transform (Cooley and Tukey, 1965) made the decomposition cheap enough to run in real time, which is why your phone can filter a voice or tune a radio on the fly. Any time you speak of "the frequencies in" a sound, an image, or a data stream, you are standing on Fourier's heat equation.