Karl Weierstrass

Meet Karl Weierstrass (1815–1897): the man mathematicians call the father of modern analysis, and the owner of one of the great late-blooming careers in the history of science. For his entire twenties and most of his thirties he was nobody in particular — an obscure small-town schoolteacher marking penmanship exercises and supervising gymnastics, publishing nothing, known to no mathematician alive. Then, in his late thirties, he sent one paper to a journal, and within a few years he held one of the most prestigious mathematics chairs in the world. Along the way he finished a job Cauchy had started, built a function so strange it scandalised his own colleagues, and quietly became the most sought-after mathematics teacher in Europe.

The son who was supposed to become a bureaucrat

Weierstrass's father was a minor Prussian customs official who had one plan for his obviously clever son: a safe, respectable career in law and public finance. So in 1834, seventeen-year-old Karl was sent off to the University of Bonn to study exactly that. He obeyed the letter of the plan and ignored its spirit completely. Rather than sit in law and finance lectures, he spent his university years fencing, drinking with the student societies, and — in whatever hours were left — secretly teaching himself mathematics from the collected works of the Norwegian mathematician Abel, working through proofs alone with no supervisor and no one to check his reasoning.

The predictable result was that he never finished his official degree at all. He came home having spent four years and his father's money on a course of study he had privately abandoned for a different one — a colossal, career-defining act of stubbornness that would have ended most people's prospects on the spot. Instead he sat a separate set of exams to qualify as a schoolteacher, and in 1841, at twenty-six, took a post at a gymnasium — a German secondary school — in the small town of Münster, later moving to similarly obscure postings at Deutsch Krone and Braunsberg. For the next decade and a half he taught teenagers not only mathematics but handwriting, geography, and gymnastics, while doing genuinely original research alone, at night, with no university library, no colleagues to argue with, and no journal reading his work — because he wasn't sending them any.

It's worth sitting with just how ordinary Weierstrass's working life looked for those fifteen-odd years. His job title was Gymnasiallehrer — the German equivalent of a secondary-school teacher — and his official duties at various points included drilling children in cursive handwriting and running physical education classes. He had no doctorate, no university post, and by his late thirties had published almost nothing.

Anyone judging him purely on his CV in, say, 1853, would have filed him as a mildly overqualified schoolmaster in provincial Prussia — not the person who was, in his spare evenings, quietly working out results in the theory of abelian functions that would soon be recognised as some of the deepest mathematics of the century. It is one of the strangest "hidden genius" stories in the whole history of science, precisely because there was no eccentric-professor mythology attached to it at the time — just a schoolteacher, correcting essays by day and doing world-class mathematics by lamplight.

One paper, and then a chair overnight

The turning point came in 1854, when Weierstrass — now thirty-nine — published a paper on abelian functions, a difficult generalisation of trigonometric and elliptic functions that several of the era's leading mathematicians had been struggling with for years. It solved a problem that had defeated better-credentialed people, and it did so with a technical power that made it obvious the author was no amateur.

The mathematical world reacted the way you'd hope a good story would go. The University of Königsberg promptly awarded Weierstrass an honorary doctorate despite his never having finished a degree. Offers of university positions followed almost immediately, and in 1856 he was appointed to a chair at the Berlin Trade Academy and then, that same year, given a professorship at the University of Berlin — one of the most prestigious mathematical addresses on the planet. A man who had spent a decade and a half marking penmanship exercises in provincial towns nobody outside Prussia had heard of was, within two years, one of the most senior mathematicians in Germany. Few careers in mathematics have ever pivoted so completely, so late, or so fast.

Finishing what Cauchy started

Weierstrass's defining project, once he reached Berlin, was rigour. Cauchy had already dragged calculus a long way from vague talk of quantities "becoming infinitely small," but even Cauchy's own definitions still leaned on intuitive, half-dynamic language — a variable "approaching" a value, almost as if time were passing while it did. Weierstrass wanted every trace of that motion-talk gone. His version of the limit is the one you still learn today, expressed purely in terms of fixed numbers and inequalities, with no "approaching" verb anywhere in sight:

\forall \varepsilon > 0 \;\; \exists \delta > 0 : \; 0 < |x - a| < \delta \;\Rightarrow\; |f(x) - L| < \varepsilon.

Read it as a challenge-and-response: however small a margin of error \varepsilon you demand around the value L, I can hand you a window of x-values, of width \delta, close enough to a that f(x) lands inside your margin every time. Nothing moves; nothing approaches anything. It is a static, checkable statement about numbers, and it finally let mathematicians prove facts about limits, continuity and derivatives with the same total certainty as a proof in geometry. This is the language behind ε–δ limits as taught in every analysis course since.

Weierstrass pushed the same insistence on precision into the idea of uniform convergence — nailing down exactly when a sequence of functions can be trusted to preserve properties like continuity in its limit, and when it can't — and, working with his students, gave the first fully rigorous construction of the real numbers themselves, so that even the objects calculus talks about, not just its arguments, rested on solid ground. By the time he was finished, the shaky, intuition-built calculus that Newton and Leibniz had left behind two centuries earlier had a foundation as solid as arithmetic.

The monster with no tangents

Then, in 1872, Weierstrass delivered the result that made him briefly infamous rather than merely respected. Every mathematician of the era — including some very good ones — had quietly assumed that a continuous curve, one you can draw without lifting your pen, must surely be smooth almost everywhere: however jagged it gets in places, there ought to be a well-defined tangent line at most points. It seemed less like a theorem than common sense.

Weierstrass built a function that broke this common sense completely — continuous everywhere, yet differentiable nowhere. Its graph never has a single break or jump, so you genuinely could trace it without lifting your pen, and yet at every point along the way, without exception, it is too jagged to have a tangent line at all. Zoom in anywhere, as far as you like, and instead of the curve straightening out and looking more and more like a line — the way any ordinary smooth curve does when you zoom in far enough — it stays exactly as jagged as it was before.

W(x) = \sum_{n=0}^{\infty} a^{n} \cos\!\big(b^{n} \pi x\big), \qquad 0 < a < 1,

with b an odd integer chosen large enough (Weierstrass's original condition was roughly ab > 1 + \tfrac{3}{2}\pi) that the ever finer wiggles piling on top of each other never smooth out, no matter how far you zoom. Each added term is a smaller, faster wiggle riding on top of the last; add infinitely many, and the result is a curve that is rough at every single scale simultaneously. Long before the word "fractal" existed, Weierstrass had drawn one — a shape that looks the same kind of jagged whether you view the whole thing or a billionth of it.

His contemporaries did not take it well. The distinguished French mathematician Charles Hermite reportedly called such pathological functions an object of "fright and horror," and other mathematicians of the era spoke of them as a "deplorable evil." The reaction says something interesting about mathematics itself: an entire community had been trusting a piece of intuition — smooth curves are smooth "almost everywhere" — that turned out to be simply false, and nobody had noticed until someone bothered to check with full rigour instead of a sketch.

It's easy to hear "a curve you can draw without lifting your pen, but that has no tangent anywhere" and file it as a mathematician's party trick — a deliberately weird counterexample built purely to be weird, with no real content behind it. That reading badly undersells what Weierstrass actually did.

The function was Weierstrass's answer to a genuine, serious, open mathematical question: does continuity imply differentiability almost everywhere? Mathematicians had strong intuitions on this — most believed the answer was surely yes, or at worst that any counterexample would have to be some contrived, artificial-looking curve with only a few bad points. Weierstrass showed the honest answer is a resounding no, and that the failure isn't a rare edge case at all — it can happen at literally every point of an otherwise perfectly continuous curve. That is not a magic trick; it is a load-bearing fact about the relationship between two of calculus's most basic concepts, and it forced the whole field to stop assuming continuity buys you more than it actually does.

The proof of how seriously it mattered is what came after: mathematicians didn't shrug the monster function off, they rebuilt their intuitions and their definitions around it. Once you accept that "continuous" and "smooth" are genuinely different ideas, you're ready for the far stranger, far rougher paths that show up throughout nature and mathematics — including, decades later, the ceaselessly jittering path of Brownian motion, the random jiggling of a pollen grain in water (and, later still, the model behind option pricing), which turns out to be continuous everywhere and differentiable nowhere for exactly the same underlying reason: infinite roughness at every scale.

The teacher Europe travelled to see

Away from his own research, Weierstrass became, by wide agreement, the finest mathematics lecturer of his generation. His courses at Berlin were legendary for their preparation — every definition stated with total precision, every proof built from the ground up with nothing hand-waved — and students travelled from across Europe and as far as the United States specifically to sit in his lecture hall. A striking number of the mathematicians who shaped the following generation passed through his classroom at some point, absorbing not just his results but his insistence that nothing be assumed which had not first been proved.

His most remarkable student never officially set foot in that lecture hall at all. Sofia Kovalevskaya, a brilliant young Russian mathematician, arrived in Berlin hoping to study under Weierstrass — only to find that the university flatly refused to admit women to its lectures. Rather than turn her away, Weierstrass simply tutored her privately, for years, meeting with her personally to cover the same material his enrolled students received in the hall. When her results proved good enough for a doctorate, he did not stop at teaching her — he used his own considerable prestige to lobby the University of Göttingen into granting her one in absentia, without the usual public examination a male candidate would have faced, making her one of the first women in Europe to earn a doctorate in mathematics. The two remained close friends and regular correspondents for the rest of Kovalevskaya's tragically short life; when she died young of influenza in 1891, Weierstrass was reportedly devastated, and is said to have burned much of their correspondence rather than let it become public.

A body that never quite kept up with the mind

Weierstrass paid a long-term price for those years of secret, exhausting late-night study in Bonn and the provincial schoolrooms that followed. From his forties onward he suffered from recurring, sometimes severe attacks of dizziness and nausea — probably connected to the punishing overwork of his hidden research years — which by the 1870s had become bad enough that he sometimes could not stand at a blackboard at all. Rather than stop lecturing, he adapted: for long stretches he dictated his famous, meticulously precise courses to a devoted student, who wrote the material on the board while Weierstrass sat and spoke, and later, when even sitting through a full lecture became difficult, he taught from a wheelchair. Students who came to Berlin expecting a frail invalid found instead a lecturer whose command of the material had lost none of its edge, whatever his body was doing that day.

He never married, and never had children of his own; for much of his life in Berlin he lived with two of his unmarried sisters, who ran his household and looked after him through the worst of his illness. It was, by the account of students who visited, a contented and settled domestic life built for undisturbed work — which, given how much work Weierstrass still managed to produce well into old age, seems to have suited him exactly.

A life in ten lines

1815 — born in Ostenfelde, Westphalia, the son of a Prussian customs official.
1834 — sent to the University of Bonn to study law and finance; secretly studies mathematics instead and leaves without a degree.
1841 — qualifies as a schoolteacher and begins over a decade teaching at provincial German gymnasiums.
1854 — publishes his breakthrough paper on abelian functions at age thirty-nine; awarded an honorary doctorate by Königsberg.
1856 — appointed to a professorship at the University of Berlin, one of the world's leading mathematical centres.
1861 — presents an early version of his fully rigorous ε–δ definitions to his Berlin students.
1870s — privately tutors Sofia Kovalevskaya, barred from Berlin's lectures by her gender, and later secures her a doctorate.
1872 — announces the Weierstrass function: continuous everywhere, differentiable nowhere.
1885 — proves the Weierstrass approximation theorem, showing any continuous function on a closed interval can be approximated as closely as you like by a polynomial.
1897 — dies in Berlin, having turned analysis from an intuition-driven art into a fully rigorous science.

It is hard to think of a career with a stranger shape: fifteen invisible years followed by four decades at the very top, built entirely on the strength of ideas worked out alone, at night, by a man everyone around him assumed was just a schoolteacher.

The full story (with far fewer jokes) is on Wikipedia: Karl Weierstrass — Wikipedia. For the monster curve, see the article on the Weierstrass function; for his most famous student, Sofia Kovalevskaya.