Bernhard Riemann

Bernhard Riemann (1826–1866) lived only thirty-nine years, was sickly for most of them, and was so painfully shy that he reportedly dreaded standing up in front of a room. And yet he reshaped more of mathematics — and, sixty years after his death, physics itself — than almost anyone who has ever lived. He gave the world the first rigorous definition of the integral, invented the geometry that Einstein needed for general relativity decades before Einstein needed it, and left behind a single throwaway guess about prime numbers that mathematicians still cannot prove is true. Quiet men can leave very loud footprints.

A pastor's son who couldn't stop calculating

Riemann was born in the village of Breselenz in the Kingdom of Hanover, the second of six children of a poor Lutheran pastor. The family had little money and worse luck: Bernhard's mother died while he was still young, and of his five siblings, most did not live to old age either — a shadow of frail health seems to have hung over the whole household, and it would eventually claim Bernhard too. As a boy he was famously withdrawn, terrified of speaking in public, and prone to nervous breakdowns under pressure — hardly the profile you would predict for someone who would one day rewrite the geometry of the universe. What he did have, from very early on, was a gift for mathematics so obvious that his schoolteachers simply ran out of things to teach him and handed him advanced textbooks to work through on his own.

He enrolled at the University of Göttingen in 1846, officially to study theology and please his father, but he could not stay away from mathematics for long and switched properly soon after. There he came under the eye of the aging Carl Friedrich Gauss — by then the most famous mathematician alive, and a man who did not hand out praise the way other professors hand out attendance marks. When Riemann submitted his 1851 doctoral dissertation, on the foundations of what we now call complex analysis, Gauss's official report called it proof of "a gloriously fertile originality" — extraordinary language from a man whose usual register ran from silence to grudging approval. Riemann had arrived.

1854: the lecture almost nobody understood

To become an unpaid lecturer at a German university, a young scholar had to deliver a Habilitation lecture — a kind of formal audition — usually on a topic the candidate proposed and the faculty selected from a shortlist. Riemann, characteristically, listed several safe topics and one wildly ambitious one almost as an afterthought: "On the Hypotheses which lie at the Foundations of Geometry." Gauss, who rarely showed interest in anything, picked that one.

Riemann had only weeks to prepare a talk generalising the whole idea of "space" itself — while also, by his own letters, suffering from exhaustion and nerves so severe that he nearly collapsed beforehand. He delivered it anyway, on 10 June 1854, to a room of professors who mostly could not follow where he was going. One person in that room could: Gauss, by then in the final year of his life, listened to a young man describe geometries of any number of dimensions, curved in ways that had no picture in ordinary experience — and was, by every account that survives, genuinely astonished. Praise from Gauss was vanishingly rare. Praise from Gauss for an idea this strange, delivered by a nervous 27-year-old who nearly didn't make it to the podium, was close to unheard of.

What Riemann had actually built was a way to describe curved spaces of any dimension using nothing but a rule, at every point, for measuring nearby distances — freeing geometry entirely from the flat page and the three dimensions of ordinary experience. It was, at the time, an answer with no question attached: a beautiful piece of pure mathematics that solved a problem nobody outside the room had yet realised they had.

It's tempting to tell the story backwards — to imagine Riemann somehow anticipating relativity, sketching out spacetime before anyone knew what spacetime was for. That gets the history exactly the wrong way round, and the truth is more interesting.

Riemann's geometry of curved spaces was pure mathematics, for its own sake, developed in 1854 with no physics attached to it at all — it sat in journals for sixty years, admired by a small circle of geometers, used by essentially nobody. Then, in the early 1910s, a young patent-office physicist named Albert Einstein was trying to describe gravity as the curving of space and time itself, and kept hitting a wall: he did not have the mathematical machinery to describe a "curved" four-dimensional spacetime precisely. As the story is usually told, it was his friend and fellow student Marcel Grossmann who pointed him toward exactly the geometry Riemann (and later Riemann's successors, especially the Italian school of Ricci-Curbastro and Levi-Civita) had already built — and Einstein had to sit down and learn a sixty-year-old branch of pure mathematics before he could finish general relativity in 1915.

So the tool did not follow the need; the need arrived decades after the tool, built by someone who never lived to see a single one of its applications. It's one of the cleanest examples in the history of science of "useless" pure mathematics turning out, much later and for reasons nobody could have predicted, to be exactly the load-bearing part of a completely different field.

Nailing down what an integral even is

Long before curved space, Riemann had already fixed something much more down-to-earth: nobody had rigorously defined what it means to find the area under a curve. Mathematicians had been "integrating" for over a century and a half using Newton's and Leibniz's calculus, but the underlying definition was informal — a sum of infinitely many infinitely thin slices, which sounds fine until someone asks you to say exactly what that means. Riemann supplied the precise recipe now taught in every calculus course: chop the region under the curve into thin vertical strips, add up their areas, and ask what happens as the strips get arbitrarily thin.

\int_a^b f(x)\,dx = \lim_{\|P\| \to 0} \sum_{i} f(x_i^{*})\,\Delta x_i.

This is the Riemann integral, and it did for integration roughly what Dedekind's cuts did for the real numbers around the very same years: it turned a picture everyone trusted into a definition precise enough to build theorems on — including, crucially, a clear answer to exactly which functions can and cannot be integrated at all.

Gluing together a new kind of surface

Riemann's doctoral work — the very dissertation that so impressed Gauss — introduced another idea that quietly rewired an entire field: the Riemann surface. Some functions of a complex number are stubbornly multi-valued — the square root of z, for instance, wants to give two different answers for every input, and the usual flat complex plane has no room to keep them both without a contradiction. Riemann's fix was almost architectural: instead of forcing the function onto a flat plane, build it a custom-shaped surface to live on instead — sheets of the plane, stacked and glued together along carefully chosen cuts, so that each of the function's values gets its own consistent home and the multi-valuedness simply resolves itself geometrically.

The effect was to fuse two subjects that had previously lived in separate rooms: complex analysis, the study of functions of complex numbers, and topology, the study of shapes that can bend and stretch but not tear. A function's Riemann surface is a genuine topological object — it has a shape, a number of "handles," a personality — and that shape turns out to control deep facts about the function living on it. It was one more instance of the pattern that runs through Riemann's whole career: take a problem that looks like algebra or analysis, find the geometric shape secretly hiding inside it, and let the geometry do the heavy lifting.

The million-dollar guess

In 1859, in a single nine-page paper — his only publication on number theory — Riemann studied a function now called the Riemann zeta function, and noticed it seemed to connect, in an unexpected way, to how the prime numbers are scattered among the ordinary counting numbers. Buried in the paper is a single, almost offhand conjecture about exactly where a certain family of the function's "zeros" must lie. Riemann himself admitted he hadn't been able to prove it, and moved on to other things.

That throwaway guess is now the Riemann Hypothesis, arguably the single most famous unsolved problem in all of mathematics. More than a century and a half of the best mathematicians alive have tried and failed to prove it; entire careers have been built on results that say, in effect, "if the Riemann Hypothesis is true, then here is what else follows." The Clay Mathematics Institute lists it as one of seven Millennium Prize Problems, with a standing one-million-dollar reward for a correct proof (or disproof). One retiring, publication-shy man, in nine pages, handed mathematics a puzzle it still hasn't solved a hundred and sixty-plus years later.

It's worth sitting with just how strange the Riemann Hypothesis's endurance really is. Fermat's Last Theorem — another legendarily hard problem, jotted in a margin in the 1600s — was finally proved in 1994, after 358 years, by Andrew Wiles working largely alone for seven years. The Riemann Hypothesis has now outlasted that entire saga and shows no sign of following it into the "solved" column. It has been checked by computer for the first many trillions of the zeros in question, and every single one lands exactly where Riemann guessed — a staggering amount of numerical evidence in its favour, and yet "checked for trillions of cases" is not the same as "proved for all cases," and mathematics cares about the difference more than almost any other subject on Earth.

David Hilbert, asked what he would want to know if he awoke after a thousand years of sleep, is said to have answered: whether the Riemann Hypothesis had been proved. It is the kind of problem that quietly humbles anyone tempted to think mathematics is a subject that gets "finished."

A perfectionist who once got caught out

Riemann's shyness was not an act. Colleagues describe him as almost unable to make small talk, prone to withdrawing from company for weeks at a stretch, and happiest alone with a problem. He worked slowly and revised endlessly, publishing a small fraction of what he actually discovered because he refused to release an argument until he trusted every step of it — which makes one episode from his career quietly instructive, because for once his usual caution slipped.

In some of his most important work on complex functions, Riemann leaned on a shortcut called the Dirichlet principle — a plausible-sounding assumption that a certain minimisation problem always has a solution, borrowed from his friend and mentor Peter Gustav Lejeune Dirichlet. It felt obviously true. Decades later, the mathematician Karl Weierstrass showed that the principle, as stated, was not generally valid at all — there are minimisation problems of exactly this shape with no solution. Riemann's results turned out to be correct regardless (David Hilbert eventually supplied a fully rigorous replacement argument), but the episode is a useful reminder even about a mind this careful: even the most rigour-obsessed mathematician of his generation once trusted a beautiful-looking shortcut a little further than the logic actually allowed. Mathematics forgave him; it does not always forgive everyone.

A short life, spent almost entirely working

Riemann's health was never good, and it worsened steadily through his thirties. He married Elise Koch in 1862, and the couple had a daughter, Ida — a brief window of ordinary happiness in a life otherwise dominated by illness and the slow grind of proving theorems under pressure. Hoping the warmer climate might help his failing lungs, Riemann travelled repeatedly to Italy in his final years, chasing sunshine the way other patients of the era chased any remedy that might buy them time.

It didn't buy enough. Bernhard Riemann died of tuberculosis on 20 July 1866 near Lake Maggiore in northern Italy, only thirty-nine years old, with his wife by his side. He had published only a modest handful of papers in his lifetime — a fraction of the output of many far less significant mathematicians — because he worked slowly, revised obsessively, and refused to publish anything before he considered it truly finished. It is one of the great "what ifs" of mathematics: a man this original, given only thirty-nine years and chronic ill health, still managed to reshape analysis, geometry and number theory each at least once. His friend Richard Dedekind, left behind with a desk full of unfinished papers, spent years afterwards helping to edit and publish Riemann's collected works, making sure the ideas of a friend who ran out of time still found their way into the world.

A life in eight lines

Very few scientists get to reshape one field. Riemann, in a life shorter than most people's careers, reshaped three — and left a fourth waiting a hundred and sixty years for anyone to finish what he started.

Read more

The full story (with far fewer jokes) is on Wikipedia: Bernhard Riemann — Wikipedia. For the 1854 lecture, see Riemannian geometry; for the unsolved problem, the Riemann Hypothesis.