Modular Forms — Definitions, Examples, and Hecke Operators

⚡ Key Takeaways

A modular form of weight $k$ for a congruence subgroup $\Gamma \subseteq \mathrm{SL}_2(\mathbb{Z})$ is a holomorphic function $f$ on the upper half-plane $\mathbb{H}$ satisfying the transformation law $f\!\left(\frac{a\tau+b}{c\tau+d}\right) = (c\tau+d)^k f(\tau)$ , plus holomorphy at the cusps. The space $M_k(\Gamma)$ is finite-dimensional; its key arithmetic structure comes from the commuting family of Hecke operators $T_n$ , whose simultaneous eigenforms have multiplicative Fourier coefficients encoding deep arithmetic. Applies across number theory, the theory of $L$ -functions, and (via Wiles) the proof of Fermat's Last Theorem. Foundational work: Eisenstein, Jacobi, Hecke (1937), Petersson.

💡 The Core Insight

A modular form is best understood as a holomorphic section of a line bundle on a modular curve. The transformation law $f(\gamma\tau) = (c\tau+d)^k f(\tau)$ is precisely the cocycle condition making $f(\tau)(d\tau)^{k/2}$ an invariant tensor. Finiteness of $\dim M_k(\Gamma)$ then follows from Riemann–Roch on the compact curve $X(\Gamma)$ . The Hecke operators arise from correspondences on this curve (averaging over sublattices of fixed index), and because these correspondences commute and respect the harmonic/ $L^2$ structure, the simultaneous eigenforms carry the arithmetic.

DefinitionThe upper half-plane and the modular group action

Let $\mathbb{H} = \{\tau \in \mathbb{C} : \mathrm{Im}(\tau) > 0\}$ . The group $\mathrm{GL}_2^+(\mathbb{R})$ of real $2\times 2$ matrices with positive determinant acts on $\mathbb{H}$ by Möbius transformations:

1

\quad \gamma \tau = \frac{a\tau + b}{c\tau + d}, \qquad \gamma = \begin{pmatrix} a & b \\ c & d \end{pmatrix}, \quad \det\gamma = ad - bc > 0.

Derivation hint: Compute $\mathrm{Im}(\gamma\tau) = \frac{\det\gamma\,\mathrm{Im}(\tau)}{|c\tau+d|^2}$ to confirm $\mathbb{H}$ is preserved. The modular group is $\mathrm{SL}_2(\mathbb{Z})$ , generated by $T = \begin{pmatrix}1&1\\0&1\end{pmatrix}$ ( $\tau\mapsto\tau+1$ ) and $S = \begin{pmatrix}0&-1\\1&0\end{pmatrix}$ ( $\tau\mapsto -1/\tau$ ). A congruence subgroup of level $N$ is any subgroup $\Gamma$ containing the principal congruence subgroup

2

\quad \Gamma(N) = \left\{ \gamma \in \mathrm{SL}_2(\mathbb{Z}) : \gamma \equiv \begin{pmatrix}1&0\\0&1\end{pmatrix} \pmod{N} \right\}.

Derivation hint: Reduction mod $N$ gives a homomorphism $\mathrm{SL}_2(\mathbb{Z})\to\mathrm{SL}_2(\mathbb{Z}/N)$ ; $\Gamma(N)$ is its kernel. The standard intermediate groups are $\Gamma_0(N)$ (lower-left entry $\equiv 0$ ) and $\Gamma_1(N)$ (additionally diagonal $\equiv 1$ ).

DefinitionThe weight-

k

slash operator and modular forms

For $\gamma = \begin{pmatrix}a&b\\c&d\end{pmatrix}\in\mathrm{GL}_2^+(\mathbb{R})$ and an integer $k$ , define the weight- $k$ slash action on functions $f:\mathbb{H}\to\mathbb{C}$ :

3

\quad (f|_k\gamma)(\tau) = (\det\gamma)^{k/2}\,(c\tau+d)^{-k}\, f(\gamma\tau).

Derivation hint: The factor $j(\gamma,\tau)=c\tau+d$ satisfies the cocycle identity $j(\gamma\delta,\tau)=j(\gamma,\delta\tau)\,j(\delta,\tau)$ , making $|_k$ a right group action. A modular form of weight $k$ for $\Gamma$ is a holomorphic $f:\mathbb{H}\to\mathbb{C}$ satisfying: 1. Invariance: $f|_k\gamma = f$ for all $\gamma\in\Gamma$ , i.e. $f(\gamma\tau) = (c\tau+d)^k f(\tau)$ . 2. Holomorphy at cusps: for every $\alpha\in\mathrm{SL}_2(\mathbb{Z})$ , $f|_k\alpha$ admits a Fourier expansion with only non-negative frequencies. For $\Gamma = \mathrm{SL}_2(\mathbb{Z})$ , since $f(\tau+1)=f(\tau)$ , we have the $q$ -expansion with $q = e^{2\pi i \tau}$ :

4

\quad f(\tau) = \sum_{n=0}^{\infty} a_n q^n, \qquad q = e^{2\pi i\tau}.

Derivation hint: Periodicity under $T$ makes $f$ a function of $q$ on the punctured disk; holomorphy at $i\infty$ kills negative powers. If $a_0 = 0$ at every cusp, $f$ is a cusp form. Write $M_k(\Gamma)$ for the space of modular forms and $S_k(\Gamma)\subseteq M_k(\Gamma)$ for cusp forms.

TheoremFinite-dimensionality and the dimension formula

For each $k$ , $M_k(\mathrm{SL}_2(\mathbb{Z}))$ is finite-dimensional. Modular forms exist only for even $k\ge 0$ , and

5

\quad \dim M_k(\mathrm{SL}_2(\mathbb{Z})) = \begin{cases} \lfloor k/12\rfloor + 1, & k\not\equiv 2 \pmod{12},\\[4pt] \lfloor k/12\rfloor, & k\equiv 2 \pmod{12}, \end{cases} \quad (k\ge 0 \text{ even}).

Derivation hint: Apply the valence formula $v_\infty(f) + \tfrac12 v_i(f) + \tfrac13 v_\rho(f) + \sum' v_P(f) = k/12$ counting zeros over the fundamental domain; orbifold points $i,\rho$ contribute fractional weights. Moreover $\dim S_k = \dim M_k - 1$ for $k\ge 4$ even (and $S_k = 0$ for $k < 12$ ).

TheoremHecke operators and eigenform multiplicativity

For each $n\ge 1$ there is a linear operator $T_n$ on $M_k(\mathrm{SL}_2(\mathbb{Z}))$ preserving $S_k$ , given on $q$ -expansions by

6

\quad (T_n f)(\tau) = \sum_{m=0}^\infty \left( \sum_{d \mid \gcd(m,n)} d^{k-1} a_{mn/d^2} \right) q^m.

Derivation hint: Sum $f|_k\gamma$ over coset representatives $\begin{pmatrix}a&b\\0&d\end{pmatrix}$ with $ad=n$ , $0\le b<d$ , then re-collect Fourier coefficients. The $T_n$ commute, satisfy $T_m T_n = \sum_{d\mid\gcd(m,n)} d^{k-1} T_{mn/d^2}$ , and are self-adjoint for the Petersson inner product

7

\quad \langle f, g\rangle = \int_{\Gamma\backslash\mathbb{H}} f(\tau)\overline{g(\tau)}\, (\mathrm{Im}\,\tau)^k \, \frac{d x\, d y}{y^2}, \qquad \tau = x+iy.

Derivation hint: The measure $y^{-2}\,dx\,dy$ is $\mathrm{SL}_2(\mathbb{R})$ -invariant; $y^k|f|^2$ is $\Gamma$ -invariant by the transformation law, so the integrand descends to the quotient. Hence $S_k$ has an orthogonal basis of simultaneous Hecke eigenforms. A normalized ( $a_1=1$ ) eigenform satisfies $T_n f = a_n f$ , so its coefficients are multiplicative: $a_{mn}=a_m a_n$ for $\gcd(m,n)=1$ , and $a_{p^{r+1}} = a_p a_{p^r} - p^{k-1} a_{p^{r-1}}$ .

ProofHecke operators commute and act multiplicatively

Step 1 — Lattice formulation. Identify weight- $k$ forms with functions $F$ on lattices $L\subset\mathbb{C}$ homogeneous of degree $-k$ : $F(\lambda L) = \lambda^{-k}F(L)$ , via $f(\tau) = F(\mathbb{Z}\tau + \mathbb{Z})$ . The transformation law [3] is exactly invariance under change of basis of $L$ .

Step 2 — Define $T_n$ as a lattice sum. Set

(T_n F)(L) = n^{k-1} \sum_{[L:L']=n} F(L'),

summing over all sublattices $L'\subseteq L$ of index $n$ . Homogeneity of degree $-k$ is preserved.

Step 3 — Coset representatives. Sublattices of index $n$ in $\mathbb{Z}^2$ correspond to matrices $\begin{pmatrix}a&b\\0&d\end{pmatrix}$ with $ad=n$ , $0\le b<d$ , up to left $\mathrm{SL}_2(\mathbb{Z})$ . Translating to $f$ yields the double-coset average

T_n f = n^{k/2 - 1}\sum_{ad=n,\,0\le b<d} f\big|_k\begin{pmatrix}a&b\\0&d\end{pmatrix},

and expanding the slash gives the coefficient formula [6].

Step 4 — Commutativity. $T_m T_n$ corresponds to summing $F$ over chains $L\supseteq L''\supseteq L'$ of indices $m$ and $n$ . Counting such chains by the Smith normal form of the total index shows the composite depends only on the product structure of sublattices, yielding the symmetric relation $T_m T_n = \sum_{d\mid\gcd(m,n)} d^{k-1}T_{mn/d^2}$ . Symmetry in $(m,n)$ forces commutativity.

Step 5 — Multiplicativity of eigenvalues. If $T_n f = a_n f$ for a normalized eigenform with $a_1 = 1$ , reading the $q^1$ -coefficient of $T_n f$ in [6] gives $a_n$ directly, and the operator relation transfers verbatim to the eigenvalues: $a_m a_n = \sum_{d\mid\gcd(m,n)} d^{k-1} a_{mn/d^2}$ . $\blacksquare$

∎

mathematician · Erich Hecke | 1887–1947

Hecke, a student of Hilbert, introduced the operators bearing his name in 1937, unifying the multiplicative structure of Fourier coefficients with the Euler-product factorization of $L$ -functions. He proved that the $L$ -function $L(f,s) = \sum a_n n^{-s}$ of an eigenform factors as $\prod_p (1 - a_p p^{-s} + p^{k-1-2s})^{-1}$ and satisfies a functional equation. His vision—that modular forms generate $L$ -functions with arithmetic meaning—became the seed of the Langlands program. Hans Petersson, working alongside, supplied the inner product [7] that makes self-adjointness, and hence the eigenbasis, rigorous.

🧠 Intuition

Think of $M_k(\Gamma)$ in three simultaneous languages. Geometrically, a weight- $k$ form is a section of $\omega^{\otimes k}$ on the modular curve $X(\Gamma)$ , where $\omega$ is the Hodge bundle of the universal elliptic curve; the cusp condition is just regularity at the boundary points. Analytically, the $q$ -expansion turns $f$ into a power series, so the whole space is a vector space of power series cut out by a symmetry. Arithmetically, the Hecke operator $T_p$ "averages $f$ over the $p+1$ ways an elliptic curve admits a degree- $p$ isogeny"—this is why $T_p$ is a correspondence and why $a_p$ counts something arithmetic (e.g. points mod $p$ on the associated motive). The finiteness is the small miracle: a fundamental domain has finite hyperbolic volume, so by Riemann–Roch there is no room for an infinite-dimensional space.

Figure

Fundamental domain of

\mathrm{SL}_2(\mathbb{Z})

acting on

\mathbb{H}

, with the cusp at

i\infty

and elliptic points

i,\rho

ExampleEisenstein series

E_4

and the value

E_4(i)

structure

Given. The normalized weight- $4$ Eisenstein series for $\mathrm{SL}_2(\mathbb{Z})$ ,

E_4(\tau) = 1 + 240\sum_{n=1}^\infty \sigma_3(n) q^n, \qquad \sigma_3(n) = \sum_{d\mid n} d^3.

Since $\dim M_4 = 1$ , $E_4$ spans $M_4$ . We verify it is a $T_2$ -eigenform and find its eigenvalue.

Step 1. Read off coefficients: $a_0 = 1$ , $a_1 = 240\sigma_3(1) = 240$ , $a_2 = 240\sigma_3(2) = 240\cdot 9 = 2160$ .

Step 2. Apply [6] with $n=2$ , $k=4$ . The $q^1$ -coefficient of $T_2 E_4$ is $\sum_{d\mid\gcd(1,2)} d^{3} a_{2/d^2} = a_2 = 2160$ .

Step 3. Since $\dim M_4 = 1$ , $T_2 E_4 = \lambda E_4$ , so matching $q^1$ coefficients: $\lambda \cdot 240 = 2160$ .

Boxed answer.

\boxed{\;\lambda = T_2\text{-eigenvalue} = \frac{2160}{240} = 9 = \sigma_3(2) = 1 + 2^3.\;}

This confirms the general fact that for Eisenstein series the Hecke eigenvalue is $a_p = 1 + p^{k-1}$ , here $1 + 2^3 = 9$ .

ExampleThe discriminant cusp form

\Delta

and Ramanujan's

\tau

Given. The weight- $12$ cusp form

\Delta(\tau) = q\prod_{n=1}^\infty (1-q^n)^{24} = \sum_{n=1}^\infty \tau(n) q^n,

with $\tau(1)=1$ , $\tau(2)=-24$ , $\tau(3)=252$ , $\tau(4)=-1472$ , $\tau(5)=4830$ , $\tau(6)=-6048$ . Since $\dim S_{12}=1$ , $\Delta$ is automatically a normalized Hecke eigenform.

Step 1 — Verify multiplicativity at $(2,3)$ . Multiplicativity predicts $\tau(6) = \tau(2)\tau(3)$ . Compute: $\tau(2)\tau(3) = (-24)(252) = -6048 = \tau(6)$ . ✓

Step 2 — Verify the prime-power recursion at $p=2$ . With $k=12$ : $\tau(4) = \tau(2)^2 - 2^{11}\tau(1) = (-24)^2 - 2048 = 576 - 2048 = -1472$ . ✓

Step 3 — Euler product / $L$ -function. Multiplicativity gives the factorization

L(\Delta,s) = \sum_{n\ge 1}\frac{\tau(n)}{n^s} = \prod_p \frac{1}{1 - \tau(p)p^{-s} + p^{11-2s}}.

Boxed answer.

\boxed{\;\tau(2)=-24,\ \tau(4)=-1472,\ \tau(6)=-6048;\quad \tau(mn)=\tau(m)\tau(n)\ \text{for}\ \gcd(m,n)=1.\;}

Deligne's bound (proved 1974) gives $|\tau(p)| \le 2 p^{11/2}$ ; e.g. $|\tau(2)|=24 \le 2\cdot 2^{5.5}\approx 90.5$ . ✓

⚠️ Common mistake

The weight must be even for $\mathrm{SL}_2(\mathbb{Z})$ . Since $-I \in \mathrm{SL}_2(\mathbb{Z})$ acts trivially on $\mathbb{H}$ but contributes the factor $(-1)^k$ in the slash, condition $f|_k(-I) = f$ forces $(-1)^k = 1$ . Odd-weight forms exist only for groups not containing $-I$ , e.g. $\Gamma_1(N)$ with $N\ge 3$ , where they require a nontrivial nebentypus character.

"Modular form" ≠ "cusp form." Eisenstein series are modular forms with $a_0 \neq 0$ ; cusp forms vanish at every cusp. The Hecke action preserves both, but the multiplicative theory's cleanest statements (Euler products with the right functional equation) belong to the cuspidal eigenforms, while Eisenstein eigenvalues are simply $\sigma_{k-1}$ -type and factor as products of Riemann zeta values.

Holomorphy at the cusp is essential. Dropping it admits weakly holomorphic and meromorphic modular forms (e.g. the $j$ -invariant $j = E_4^3/\Delta$ with a pole at $i\infty$ ), which are not in $M_k$ and behave very differently dimensionally.

Counterexample

A function satisfying the transformation law but failing holomorphy at the cusp. The non-holomorphic Eisenstein series

E_k^*(\tau) \ \text{(real-analytic, weight } k)

and, more sharply, the weight- $2$ quasimodular $E_2(\tau) = 1 - 24\sum_{n}\sigma_1(n)q^n$ is not modular: under $S$ it picks up an anomalous term,

E_2(-1/\tau) = \tau^2 E_2(\tau) + \frac{12\tau}{2\pi i}.

The extra $\frac{12\tau}{2\pi i}$ shows $E_2 \notin M_2$ (indeed $M_2(\mathrm{SL}_2(\mathbb{Z}))=0$ , consistent with [5]). Only the non-holomorphic completion $E_2^*(\tau) = E_2(\tau) - \frac{3}{\pi\,\mathrm{Im}\,\tau}$ transforms correctly—at the cost of holomorphy. This is the prototype of a quasimodular/almost-holomorphic form.

connection

Modular forms sit at a crossroads. Via Elliptic Curves over $\mathbb{Q}$ , the modularity theorem associates to each curve a weight- $2$ eigenform with $a_p = p+1-\#E(\mathbb{F}_p)$ , linking $T_p$ -eigenvalues to point counts. Through Galois Representations, Deligne attached to each eigenform a 2-dimensional $\ell$ -adic representation whose Frobenius traces are the $a_p$ , embedding modular forms in the Langlands Program. The Petersson inner product [7] makes $S_k$ a unitary $\mathrm{SL}_2(\mathbb{R})$ -representation-theoretic object, connecting to Automorphic Forms on adele groups, where Hecke operators become local components of the global Hecke algebra. The $L$ -functions $L(f,s)$ tie modular forms to Analytic Number Theory and the Riemann Zeta Function (whose values appear in Eisenstein coefficients). Finally, the bundle interpretation links them to Riemann–Roch Theorem on the curve $X(\Gamma)$ .

🔬 Where this appears

⏳ History

- 1750s–1827. Euler and then Jacobi study theta functions and elliptic integrals; Jacobi's Fundamenta Nova (1829) develops the modular equation and identities that are implicitly modular-form statements.

- 1847. Eisenstein introduces the series

G_k(\tau) = \sum_{(m,n)\neq 0}(m\tau+n)^{-k}

, the first explicit examples of modular forms.

- 1877. Dedekind studies the

\eta

-function

\eta(\tau)=q^{1/24}\prod(1-q^n)

, the weight-

1/2

building block of

\Delta = \eta^{24}

.

- 1890s. Klein and Fricke systematize the theory of modular functions and the

j

-invariant; Poincaré develops automorphic functions for general Fuchsian groups.

- 1916. Ramanujan introduces

\tau(n)

, conjectures its multiplicativity and the bound

|\tau(p)|\le 2p^{11/2}

, anticipating both Hecke and Deligne.

- 1925–1932. Hecke and Petersson lay the analytic foundations; Petersson introduces the inner product [7].

- 1937. Hecke defines the operators

T_n

, proves multiplicativity of eigenform coefficients and the Euler-product factorization of

L(f,s)

.

- 1968–1974. Deligne constructs

\ell

-adic Galois representations attached to eigenforms (1968–71) and, using the Weil conjectures, proves the Ramanujan–Petersson bound

|a_p|\le 2p^{(k-1)/2}

(1974).

- 1994–1995. Wiles and Taylor–Wiles prove modularity for semistable elliptic curves, completing the proof of Fermat's Last Theorem; full modularity (Breuil–Conrad–Diamond–Taylor) follows in 2001.

open-problem

Lehmer's conjecture (1947) asks whether $\tau(n)\neq 0$ for all $n\ge 1$ —equivalently whether the Hecke eigenform $\Delta$ ever has a vanishing coefficient. It remains open despite verification for enormous ranges. More broadly, the precise distribution of normalized eigenvalues $a_p/p^{(k-1)/2}\in[-2,2]$ obeys the Sato–Tate law (a theorem for non-CM holomorphic eigenforms, Barnet-Lamb–Geraghty–Harris–Taylor, 2011), but effective error terms and analogues for Maass forms involve open Ramanujan-type conjectures. For Maass (non-holomorphic) forms the Ramanujan–Petersson conjecture $|a_p|\le 2p^{(k-1)/2}$ is still unproven in general; the best bounds (Kim–Sarnak) fall short. The structure of Hecke algebras at the boundary—half-integral weight, the Shimura correspondence, and $p$ -adic / overconvergent modular forms (Coleman, Hida theory)—pushes the definition far beyond the classical holomorphic case and remains an active frontier.

Modular Forms — Definitions, Examples, and Hecke Operators

Formal Statement

Why It Holds

Building Intuition

Worked Examples

Common Misconceptions

Connections

Real-World Applications

Historical Development

Subtleties & Limits

See Also