How a margin note drove centuries of obsession and one secret proof

1

The organizing taunt

One unproven amateur claim can marshal centuries of professional effort more effectively than any funded research program.

Fermat, a French lawyer working for his own pleasure, left a note in Diophantus saying no three positive integers satisfy the equation once the exponent rises past two, adding that he possessed a proof too large for the margin. He never published it. After his death the sentence hardened into a standing challenge aimed at every later reader.

Singh opens here because the scene frames the book’s central uncertainty: right, bluffing, or holding a proof that would later prove flawed. The older Pythagorean equation has infinitely many whole-number solutions; the claim is that every higher exponent destroys that possibility for all time. That forever is the difficulty. Checking hundreds of cases proves nothing; a single counterexample anywhere in infinity collapses the whole statement.

Physicists and biologists can treat strong evidence as knowledge. Mathematicians, Singh stresses, demand a proof that closes every exception. The note became “Last” not because it was proved last among Fermat’s claims, but because it was the last still unresolved while other branches grew up around it, sometimes borrowing tools built only to attack it.

In practice: The familiar triple 3, 4, 5 works for squares, yet the same pattern is said to fail for every higher exponent with no exceptions left open.

2

Sacred whole numbers

The theorem descends directly from the oldest Western number-theory obsession: the search for perfect numerical harmony.

Sixth-century-BC Pythagoreans treated numbers as sacred and believed whole-number relations underlay music, cosmos, and reality itself. Integer triples satisfying the square equation fascinated them as visible order in nature. Fermat was not inventing a curiosity; he was pressing a fault line the Greeks had already found.

Diophantus, centuries later in Alexandria, catalogued such problems, and it was in the margin of that specific text that the famous note appeared. Singh also shows Fermat’s genuine rigor on a narrower cousin: he proved by infinite descent that no whole-number solution exists when the exponent is exactly four. That success is why later mathematicians refused to dismiss the wider claim as bravado.

If a problem this old, this easy to state, and this heavily worked could still resist for centuries, direct attack had clear limits. An eventual solution would need ideas Fermat could not have imagined.

In practice: Fermat’s own descent proof for fourth powers showed he possessed real methods, not merely a flamboyant margin boast.

3

Talent without the tool

Repeated failure by the greatest mathematicians shows raw talent is not the same resource as the right conceptual tool at the right time.

Euler settled the exponent-three case by extending Fermat’s methods. Sophie Germain, barred from formal education and forced to submit under a male pseudonym, built a broad approach that covered many exponents at once. Legendre and Dirichlet pushed further; Kummer invented ideal numbers to rescue factorization when ordinary arithmetic broke down.

Each advance still left the infinite general case untouched. Progress accumulated in fragments across generations, often from people solving adjacent problems without knowing they were forging weapons for this one. The record quietly shows structural exclusion shaping credit and access, especially in Germain’s path, without heavy moralizing.

By the twentieth century brute computation had checked enormous finite lists of exponents. A large finite list still says nothing about infinity. The community largely treated the problem as unsolvable with existing methods. Great problems, Singh insists, do not fall to isolated flashes; they wait on concepts that may not yet exist.

In practice: Kummer’s ideal numbers handled breakdowns in ordinary factorization yet still could not finish the unrestricted theorem.

4

The sideways bridge

The path arrived from an unrelated field, proving deep problems are often solved sideways rather than head-on.

Goro Shimura and Yutaka Taniyama proposed that every elliptic curve—smooth shapes from cubic equations—is secretly tied to a modular form, a highly symmetric object from a distant corner of complex analysis. The claim looked so strange it stayed speculative for decades.

Gerhard Frey later showed that a Fermat counterexample would let someone build an elliptic curve with impossible properties, one that could match no modular form. Kenneth Ribet made the link rigorous. The chain locked: if the conjecture held for the right curves, Fermat’s theorem had to hold, because a counterexample would break the modular connection.

An isolated riddle became a corollary of a much larger structural claim already under active study. A proof of even a limited case of the conjecture would finish the old job.

In practice: Frey’s constructed curve from a hypothetical Fermat solution would be non-modular, so proving modularity for that class kills every such solution.

5

The attic years

A problem of this size forced abandonment of mathematics’ normal open rhythm for prolonged isolated concentration.

Wiles had been seized by the theorem since finding it in a library book as a child. Once Ribet made the modular bridge explicit, he saw that proving the conjecture for semistable elliptic curves would yield Fermat. He told almost no one and worked largely alone, only rarely asking colleagues narrow technical questions that hid the larger aim.

Normal progress runs on publication, conferences, and fast circulation of partial results. He broke that model from fear that a faster rival or a bigger team would claim the childhood prize. Techniques from Galois representations, modular forms, and Iwasawa theory were combined into a proof of unprecedented density.

Singh dwells on dead ends and stretches when total failure seemed certain. Sustained single-person secrecy is rare in modern collaborative mathematics; willingness to sacrifice ordinary professional life becomes part of what the subject can demand.

In practice: Colleagues heard only isolated technical queries; the seven-year goal stayed hidden until the Cambridge lectures.

6

Announced is not verified

A public proof is unfinished until it survives scrutiny; the gap between announced and verified can equal the gap between conjecture and proof.

Three lectures in Cambridge revealed the secret work to a stunned audience and then to global media. The atmosphere felt electric; one of the oldest problems seemed to have fallen.

Referees working through the roughly 200-page manuscript found trouble. Nick Katz isolated a subtle gap in the Euler-system step used to cover all needed cases. The hole did not obviously kill the argument, yet it could not be patched by inspection.

Nearly a year of repair followed, with former student Richard Taylor, under the real fear that the entire effort might collapse. A different route—Iwasawa theory joined to Kolyvagin-Flach methods—closed the hole. Certainty, Singh underlines, is earned through adversarial checking, not the prover’s confidence. The private crisis after the public triumph becomes the book’s true climax.

In practice: Katz’s reading of the Euler-system construction forced a substantial rebuilding rather than a quick fix before the corrected proof could stand.

7

Tools outlive the trophy

The 358-year problem was worth chasing not for applications, which are negligible, but for what the pursuit forced mathematics to build.

No engineering or physics result depends on knowing whether the equation has solutions past squares. The obvious skeptical question is why brilliant people spent careers on it. Singh’s answer: the theorem acted as a forcing function.

Attempts produced ideal numbers, advances in algebraic number theory, and ultimately confirmation of the modularity claim for a broad class of curves—work with far wider consequences that feeds the Langlands program. The statement’s middle-school simplicity also made it a public symbol, drawing amateurs, cranks, and professionals in a way denser conjectures never could.

Almost everyone already suspected the claim was true. Value lay in the tools invented in the chase, tools that outlived the specific question and reshaped the surrounding field.

In practice: Modularity results demanded by the proof now sit inside one of modern number theory’s most active research agendas.