The Taniyama–Shimura Conjecture¶

Summary

The Taniyama–Shimura conjecture asserts that every elliptic curve over $\mathbb{Q}$ is modular – that the $L$-series of every such curve coincides with the $L$-series of a modular form. This bridge between geometry and analysis is the key to Wiles' proof of Fermat's Last Theorem.

Prerequisites¶

Elliptic Curves – basic concepts: Weierstraß equation, group structure, reduction modulo $p$
Modular Forms – modular forms, newforms, $q$-expansion, Hecke operators

1. Two Separate Worlds¶

Twentieth-century mathematics knew two seemingly independent domains, both possessing deep structure – but at first glance having nothing to do with each other.

The world of elliptic curves¶

An elliptic curve over $\mathbb{Q}$ is (in simplified form) the solution set of an equation of the form

\[ E: \quad y^2 = x^3 + ax + b, \qquad a, b \in \mathbb{Q}, \quad 4a^3 + 27b^2 \neq 0. \]

It lives in algebraic geometry and number theory: one studies its rational points, its structure as an abelian group, its reduction modulo primes. For each prime $p$, one counts the points of the reduced curve over $\mathbb{F}_p$ and defines

\[ a_p(E) = p - \#E(\mathbb{F}_p), \]

where $\#E(\mathbb{F}_p)$ is the number of points (including the point at infinity) on the reduced curve. These numbers $a_p$ encode the local arithmetic of the curve at each prime.

The world of modular forms¶

A modular form of weight $k$ and level $N$ is a holomorphic function

\[ f: \mathcal{H} \to \mathbb{C} \]

on the upper half-plane $\mathcal{H} = \{z \in \mathbb{C} : \text{Im}(z) > 0\}$ that transforms in a prescribed way under the action of the congruence subgroup $\Gamma_0(N) \subset \text{SL}_2(\mathbb{Z})$. It lives in complex analysis and has a Fourier expansion (also called a $q$-expansion):

\[ f(z) = \sum_{n=0}^{\infty} b_n q^n, \qquad q = e^{2\pi i z}. \]

The coefficients $b_n$ encode the structure of the modular form. Particularly important are the newforms – normalised Hecke eigenforms that are "irreducible" in the sense that they do not arise from lower level.

Why separate?¶

Elliptic curves belong to algebraic geometry and number theory. Modular forms belong to complex analysis and representation theory. Their tools, their intuitions, their languages – everything seems different. That a deep connection might exist between these worlds was scarcely imaginable until the mid-20th century.

2. The $L$-Series Bridge¶

The key to the connection lies in the $L$-series – analytic objects that equip both worlds with a common language.

The $L$-series of an elliptic curve¶

For an elliptic curve $E/\mathbb{Q}$, one defines the $L$-series as an Euler product:

\[ L(E, s) = \prod_{p \text{ prime}} L_p(E, s)^{-1}, \]

where the local factors for primes of good reduction take the form:

\[ L_p(E, s) = 1 - a_p p^{-s} + p^{1-2s}. \]

The $a_p$ are precisely the point-counting coefficients defined above. For the finitely many primes of bad reduction (which make up the conductor $N_E$ of $E$), the local factor is simpler.

The $L$-series of a modular form¶

For a newform $f = \sum b_n q^n$ of weight 2 and level $N$, one defines analogously:

\[ L(f, s) = \sum_{n=1}^{\infty} b_n n^{-s} = \prod_{p \text{ prime}} \left(1 - b_p p^{-s} + p^{1-2s}\right)^{-1}. \]

The bridge¶

The central observation is: both $L$-series have exactly the same structure. If there exists a newform $f$ of weight 2 with

\[ a_p(E) = b_p(f) \quad \text{for (almost) all primes } p, \]

then the $L$-series agree: $L(E, s) = L(f, s)$. In this case one says: $E$ is modular, and $f$ is the modular form associated to $E$.

3. Taniyama and Shimura – The Conjecture and Its History¶

The Tokyo Symposium (1955)¶

In September 1955, an international symposium on algebraic number theory took place in Tokyo. There the young Japanese mathematician Yutaka Taniyama (1927–1958) formulated a series of problems suggesting a connection between elliptic curves and modular forms. His questions were somewhat vaguely stated, but the core was revolutionary: the $L$-series of elliptic curves should agree with those of modular forms.

Taniyama's colleague Goro Shimura (1930–2019) made the conjecture precise in the following years and supported it with calculations and theoretical arguments. In the Western literature, the conjecture is therefore often called the Taniyama–Shimura conjecture (TSC), sometimes also the Taniyama–Shimura–Weil conjecture, since André Weil made an important contribution to its precise formulation.

The tragic history¶

Taniyama took his own life in 1958 at the age of only 31 – for reasons that remain not fully understood to this day. His mathematical vision, however, survived him and became one of the most influential conjectures of the 20th century.

The conjecture, precisely stated¶

Taniyama–Shimura Conjecture (Modularity Conjecture)

Every elliptic curve $E$ over $\mathbb{Q}$ is modular: there exists a newform $f$ of weight 2 and level $N_E$ (the conductor of $E$) with $$ L(E, s) = L(f, s). $$ Equivalently: $a_p(E) = b_p(f)$ for all primes $p$ of good reduction.

Why did people believe it?¶

First, there was numerical evidence: for many explicitly computed elliptic curves, matching modular forms could be found, and the coefficients agreed – as far as one could compute.

Then there were structural arguments: the functional equation of the $L$-series of a modular form was known. If the $L$-series of an elliptic curve satisfies the same functional equation (as suggested by the work of Weil), then a connection seems likely.

Finally, there was the philosophical conviction underlying the Langlands programme: between automorphic forms (to which modular forms belong) and Galois representations (which elliptic curves yield), there should be a systematic correspondence.

4. What "Modular" Means – An Example¶

Consider the elliptic curve

\[ E: \quad y^2 = x^3 - x. \]

This is a curve with conductor $N_E = 32$. We compute the coefficients $a_p$ by counting points modulo small primes:

$p$	$\#E(\mathbb{F}_p)$	$a_p = p - \#E(\mathbb{F}_p)$
3	4	$-1$
5	4	$1$
7	8	$-1$
11	12	$-1$
13	12	$1$

Now we look for a modular form $f$ of weight 2 and level 32 with the same coefficients. Indeed, there is exactly one such newform, and its $q$-expansion begins with:

\[ f(q) = q - q^3 + q^5 - q^7 - q^{11} + q^{13} + \cdots \]

The coefficients $b_3 = -1$, $b_5 = 1$, $b_7 = -1$, $b_{11} = -1$, $b_{13} = 1$ agree exactly with the $a_p$. The curve $y^2 = x^3 - x$ is modular.

This example illustrates the conjecture concretely: the arithmetic information of the curve (point counting over finite fields) is exactly mirrored by an analytic object (a modular form).

5. Why the TSC Is So Powerful¶

The Taniyama–Shimura conjecture is not merely an observation about individual examples – it is a universal statement about all elliptic curves over $\mathbb{Q}$:

Infinitely many curves, one conjecture¶

There are infinitely many non-isomorphic elliptic curves over $\mathbb{Q}$, parametrised by the coefficients $a$ and $b$. For every single one, the TSC asserts the existence of a matching modular form. This is a breathtakingly strong statement.

From geometry to analysis¶

The TSC translates a geometric-arithmetic problem (structure of an elliptic curve) into an analytic problem (existence of a modular form). Since modular forms are well-understood objects – with a rich theory of Hecke operators, $L$-series, and functional equations – the modularity of a curve immediately unlocks a wealth of analytic tools.

Consequences of modularity¶

If $E$ is modular, it follows automatically that:

Analytic continuation: $L(E, s)$ admits analytic continuation to all of $\mathbb{C}$.
Functional equation: $L(E, s)$ satisfies a functional equation relating $s$ and $2-s$.
BSD conjecture: The order of the zero of $L(E, s)$ at $s = 1$ should equal the rank of $E(\mathbb{Q})$ (Birch and Swinnerton-Dyer).

Before Wiles' proof, even the analytic continuation of $L(E, s)$ was known only for individual curves – not for all.

6. The Semistable Version¶

Semistable elliptic curves¶

An elliptic curve $E/\mathbb{Q}$ is called semistable if at every prime $p$ it has either good or multiplicative (not additive) reduction. Geometrically, this means: at primes of bad reduction, the curve has at most an ordinary double point (a "self-crossing"), but no cusp.

The class of semistable curves is large enough to contain the Frey curve – this is decisive for the application to Fermat's Last Theorem.

Wiles' Theorem (1995)¶

Theorem (Wiles, Taylor–Wiles, 1995)

Every semistable elliptic curve over $\mathbb{Q}$ is modular.

Andrew Wiles proved this statement in his groundbreaking paper "Modular elliptic curves and Fermat's Last Theorem" (Annals of Mathematics, 1995), together with the companion article by Richard Taylor and Andrew Wiles.

The proof of the full Taniyama–Shimura conjecture – for all elliptic curves, not just the semistable ones – was achieved only in 2001 by Breuil, Conrad, Diamond, and Taylor, building on Wiles' methods.

Why does the semistable version suffice for FLT?¶

The Frey curve, constructed from a hypothetical FLT solution, is semistable. Therefore the semistable version of the TSC suffices to prove Fermat's Last Theorem – nothing more was needed. The details of this reduction are the subject of the next article.

7. From TSC to FLT – Preview of Frey's Argument¶

The logical chain from the TSC to Fermat's Last Theorem can be summarised concisely:

Suppose there existed a non-trivial solution $a^p + b^p = c^p$ for a prime $p \geq 5$.

Frey (1985): Construct the elliptic curve $E: y^2 = x(x - a^p)(x + b^p)$.
Frey/Serre: This curve is semistable but has such an extreme discriminant that it is "too exotic" to be modular.
Ribet (1986): Proves that $E$ indeed cannot be modular (level-lowering theorem).
Wiles (1995): Proves that every semistable curve is modular.
Contradiction: $E$ is semistable (hence modular by Wiles), but not modular (by Ribet). Therefore the solution cannot exist.

\[ \boxed{\text{FLT solution} \xrightarrow{\text{Frey}} E \xrightarrow{\text{Ribet}} \text{not modular} \xleftarrow{\text{contradiction}} \xrightarrow{\text{Wiles}} \text{modular}} \]

This beautiful proof by contradiction – which proves a conjecture in number theory via a detour through algebraic geometry and complex analysis – is one of the great intellectual masterpieces of mathematics.

Outlook¶

This article has presented the Taniyama–Shimura conjecture as a bridge between elliptic curves and modular forms. In the next article, we dive deeper:

Article	Topic
02 – Frey's Idea and Ribet's Theorem	How an FLT solution leads to the "impossible" Frey curve
03 – Galois Representations	How Wiles translates modularity into the language of representations

Sources¶

Nigel Boston: The Proof of Fermat's Last Theorem (2003), Chapter 8 – Modularity and the TSC
Andrew Wiles: Modular elliptic curves and Fermat's Last Theorem, Annals of Mathematics 141 (1995), §1–2
Fred Diamond, Jerry Shurman: A First Course in Modular Forms, Springer (2005) – Detailed exposition of the TSC and its proofs
Barry Mazur: Number Theory as Gadfly, The American Mathematical Monthly 98 (1991) – Motivation of the conjecture

\(p\)	\(\#E(\mathbb{F}_p)\)	\(a_p = p - \#E(\mathbb{F}_p)\)
3	4	\(-1\)
5	4	\(1\)
7	8	\(-1\)
11	12	\(-1\)
13	12	\(1\)