Langlands Program Explained: A Simple Worked Example

Quick answer: The Langlands program is a network of conjectures connecting number theory with highly symmetric functions called automorphic forms. Its central correspondences relate arithmetic symmetries, encoded by Galois representations, to automorphic data. Robert Langlands proposed the foundational ideas in 1967. Important cases are proved, including the modularity results used to prove Fermat’s Last Theorem, but the full program remains open.

Mathematicians sometimes call the Langlands program a “grand unified theory of mathematics.” You can see the idea with remainders and elementary algebra. Below we count solutions of an equation modulo small primes, then calculate coefficients of a completely different object. The surprise is that the two calculations produce the same numbers.

A Check You Can Do by Hand

Take the equation

\[ y^2 + y = x^3 – x^2. \]

Instead of ordinary numbers, work “mod \(p\)”: use only the remainders \(0, 1, \dots, p-1\) and wrap around, like a clock (see our modular arithmetic calculator). Count the pairs \((x, y)\) that work and call the count \(N_p\). Then compute \(a_p = p – N_p\).

Now take a totally different object, an infinite product:

\[ q\,(1-q)^2(1-q^{11})^2\,(1-q^2)^2(1-q^{22})^2\,(1-q^3)^2(1-q^{33})^2 \cdots = q – 2q^2 – q^3 + 2q^4 + \cdots \]

Prime \(p\)Solutions \(N_p\)\(a_p = p – N_p\)Coefficient of \(q^p\) in the product
24\(-2\)\(-2\)
34\(-1\)\(-1\)
54\(1\)\(1\)
79\(-2\)\(-2\)

For \(p = 2\): \(y^2 + y = y(y+1)\) is always even, and \(x^3 – x^2 = x^2(x-1)\) is always even, so all four pairs work. For \(p = 3\): the left side takes values 0, 2, 0 for \(y = 0, 1, 2\), and the right side takes values 0, 0, 1 for \(x = 0, 1, 2\), giving 2 + 2 + 0 = 4 solutions. Try \(p = 5\) yourself as an exercise; you should find 4 solutions.

Checking the product side. Here \(q\) is a bookkeeping variable: the coefficient of \(q^2\) is the number multiplying \(q^2\), not the value you get by substituting 2. For the coefficients of \(q^2\) and \(q^3\), only two factors matter: \((1-q)^2 = 1 – 2q + q^2\) and \((1-q^2)^2 = 1 – 2q^2 + q^4\). Multiply them and then by \(q\): \(q – 2q^2 – q^3 + \cdots\). Every other factor only changes terms of degree 4 or higher. So the coefficients \(-2\) and \(-1\) come straight from the product, matching the counts mod 2 and mod 3.

It is not a coincidence. The equation is an elliptic curve (LMFDB label 11.a3), the product is a modular form, and the modularity theorem says every elliptic curve defined over the rational numbers has a corresponding modular form; this one happens to have a convenient product formula. A few matching primes illustrate the theorem; they do not prove it.

What Is the Langlands Program?

In January 1967, Robert Langlands, then a young professor at Princeton, sent a 17-page handwritten letter to André Weil. He proposed that matchings like this are part of a far larger pattern. Over the next decades the proposal grew into a network of precise conjectures.

TermBeginner’s explanation
Galois representationA way to encode the symmetries of number-theoretic equations using matrices.
Modular formA special function obeying precise symmetry rules; above we saw one through its coefficients.
Automorphic formThe broad family of highly symmetric functions that includes modular forms.
FunctorialityPredictions about transferring automorphic information between different symmetry groups.
Local vs. globalLocal questions look at one prime at a time; global questions concern the whole number system at once.

What Has Been Proved?

YearResultWho
1994 (published 1995)Semistable elliptic curves over the rationals are modular, enough for Fermat’s Last TheoremAndrew Wiles, with Richard Taylor
2001Every elliptic curve over the rationals is modularBreuil, Conrad, Diamond and Taylor
2002 (publication)Langlands correspondence for \(GL_n\) over function fieldsLaurent Lafforgue (Fields Medal)
2008 (preprint)The Fundamental Lemma, a major enabling theoremNgô Bảo Châu (Fields Medal 2010)
2024What the geometric Langlands proof established: the global, unramified categorical correspondence in characteristic zeroGaitsgory, Raskin and seven collaborators

Langlands received the Abel Prize in 2018 for the program.

Why Is It Called a Grand Unified Theory?

Because it links number theory, harmonic analysis, representation theory and geometry, and because hard problems in one area become tractable after translation into another. Fermat’s Last Theorem is the famous example. It says \(a^n + b^n = c^n\) has no solution in positive integers when \(n > 2\). Work by Gerhard Frey, Jean-Pierre Serre and Ken Ribet showed that a hypothetical counterexample would produce a semistable elliptic curve that could not be modular. Wiles then proved that semistable elliptic curves over the rationals are modular, ruling the counterexample out. Mathematician Edward Frenkel popularized the “grand unified theory” phrase, and the geometric version even appears in quantum field theory through work of Kapustin and Witten.

What Is Still Open?

The program remains open in general. Major goals include broader correspondences over number fields and general functoriality. Many important cases are known, but they do not yet form a complete theory. The 2024 geometric breakthrough settles a major correspondence in a related geometric setting, strengthening the program while leaving the arithmetic questions open.

Common Misconceptions

“Langlands is one conjecture.” It is a family of related conjectures, local and global, arithmetic and geometric.

“Proving geometric Langlands proved Langlands.” No. The geometric version is a parallel world; the arithmetic version is still largely open.

Watch: The Langlands Program in Video

If you learn better by watching, these videos are worth your time.

The Biggest Project in Modern Mathematics (Quanta Magazine). An animated introduction that uses the same kind of clock-arithmetic example.
The Langlands Program (Numberphile). Alex Kontorovich explains the big picture.
Mysteries of Math and the Langlands Program, Episode 4 (Edward Frenkel). Part of a series: Frenkel on why the program is called a grand unified theory.

Check Your Understanding

Five quick questions to see how much of this has stuck. Pick an answer and press Submit; you will see a short explanation either way.

Answer each question and press Submit to see your result right away.

1. Who proposed the Langlands program?
Found a bug? Let us know!
2. Working mod 3, the equation has 4 solutions. What is a₃ = p − N_p?
Found a bug? Let us know!
3. Which famous theorem was proved using a case of the Langlands program?
Found a bug? Let us know!
4. Do a few matching coefficients prove the modularity theorem?
Found a bug? Let us know!
5. What does the 2024 geometric proof leave unresolved?
Found a bug? Let us know!

Frequently Asked Questions

What is the Langlands program in simple terms?

It is a set of conjectures predicting a dictionary between number theory (whole-number equations and their symmetries) and harmonic analysis (automorphic forms). Problems that are hard on one side can become easier after translation to the other.

Has the Langlands program been proved?

Only in part. Major cases are proved, including the modularity theorem behind Fermat’s Last Theorem, the function field correspondence for GL(n), and the geometric Langlands conjecture in 2024. The full arithmetic program is still open.

Why is the Langlands program called a grand unified theory of mathematics?

Because it connects several large areas of mathematics, including number theory, representation theory, analysis and geometry, and predicts that they describe the same underlying structures.

How is the Langlands program related to Fermat’s Last Theorem?

Frey, Serre and Ribet showed that a counterexample to Fermat would give a semistable elliptic curve that is not modular. Andrew Wiles proved that all semistable elliptic curves over the rationals are modular, a Langlands-type statement, so no counterexample exists.

Can you give a simple example of the Langlands program?

Count the solutions of y² + y = x³ − x² modulo a prime p and subtract from p. For p = 2, 3, 5, 7 you get −2, −1, 1, −2, exactly the coefficients of q^p in the product q(1−q)²(1−q¹¹)²(1−q²)²(1−q²²)²…, a modular form.

Keep Exploring

Sources and Further Reading

Related to This Article

What people say about "Langlands Program Explained: A Simple Worked Example - Effortless Math"?

No one replied yet.

Leave a Reply