Worked solution: Wiles's modularity proof via the Taylor–Wiles method (1994)

Step 5 of 11: The Taniyama–Shimura–Weil conjecture: every elliptic curve over

\mathbb{Q}

is modular

Previous step Next step

In plain words

At first glance, elliptic curves (cubic equations with rational points) and modular forms (hyper-symmetric functions on the upper half-plane) live on different planets. Finding even one elliptic curve whose point counts match the Fourier coefficients of a modular form looks like a coincidence.

In the 1950s and 1960s, Yutaka Taniyama, Goro Shimura, and André Weil conjectured that it is never a coincidence: every single elliptic curve defined over $\mathbb{Q}$ has a matching modular form of level equal to its conductor $N$ . For decades this bold bridge between algebra and analysis was considered out of reach.

\forall\, E/\mathbb{Q} \text{ of conductor } N, \quad \exists\, X_0(N) \twoheadrightarrow E \iff L(E, s) = L(f, s)

Detailed analysis

As Wiles recounts in his opening paragraph (Wiles 1995, p. 443; see also Qiu et al. 2025, Section 2.2), the conjecture grew out of a problem posed by Yutaka Taniyama in 1955 and refined by Goro Shimura in the 1950s and 1960s. In 1967, André Weil proved that if the $L$ -series of an elliptic curve and its twists satisfy the expected functional equations, the curve must come from a modular form of level equal to its conductor $N$ , giving strong conceptual support to the Taniyama–Shimura–Weil (TSW) conjecture.

Precisely, the conjecture asserts that every elliptic curve $E/\mathbb{Q}$ of conductor $N$ admits a nonconstant rational map from the modular curve $X_0(N) \twoheadrightarrow E$ , or equivalently corresponds to a weight- $2$ , level- $N$ Hecke eigenform $f$ with $L(E, s) = L(f, s)$ . Two elliptic curves over $\mathbb{Q}$ with the same $j$ -invariant are simultaneously modular or non-modular, and prior to Wiles's work in 1995, only finitely many $j$ -invariants were proven to be modular (Wiles 1995, p. 443).

Until 1985, the TSW conjecture belonged to mainstream arithmetic geometry and had no known link to Fermat's equation. Frey's construction suddenly tied the two together: if TSW holds for semistable curves, then the hypothetical Frey curve $E_{a,b,c}$ must be modular of level $N = \operatorname{rad}(abc)$ .

Terms in this step

Modular curve $X_0(N)$: The compactified Riemann surface obtained as the quotient of the upper half-plane by the congruence subgroup $\Gamma_0(N)$ , defined as an algebraic curve over $\mathbb{Q}$ . An elliptic curve $E/\mathbb{Q}$ is modular if there is a nonconstant map $X_0(N) \twoheadrightarrow E$ over $\mathbb{Q}$ .
$j$ -invariant: A rational number computed from the coefficients of an elliptic curve that classifies the curve up to isomorphism over $\overline{\mathbb{Q}}$ . If one curve with a given $j$ -invariant is modular, every curve with that $j$ -invariant is modular (Wiles 1995, p. 443).

Knowledge used in this step

Modular forms Arithmetic geometry