21.04.01 · number-theory / modular-forms

Modular Forms on $SL_{2} (Z)$

shipped3 tiersLean: partial

Anchor (Master): Diamond-Shurman *A First Course in Modular Forms* §§1.1-1.5 + §§3.5-3.7 (full theory including dimension formula, valence formula, $L$-function attached to a cusp form); Serre *A Course in Arithmetic* Ch. VII §§1-4 (Hecke-style proof of $M_*(\mathrm{SL}_2(\mathbb{Z})) = \mathbb{C}[E_4, E_6]$ via the valence formula); Manin-Panchishkin *Introduction to Modern Number Theory* (Springer EMS 49, 2nd ed. 2005) Ch. 6 §§1-3 (modular forms in the encyclopaedic survey context, Hecke's lattice viewpoint, the bridge to $L$-functions); Hecke 1936-37 *Math. Ann.* 112 + 114 (originator papers: definition of modular forms with multiplier system, the Hecke correspondence between modular forms and Dirichlet series, the ring structure of $M_*(\mathrm{SL}_2(\mathbb{Z}))$); Shimura *Introduction to the Arithmetic Theory of Automorphic Functions* (Princeton Iwanami 1971) §§2.1-2.7 (modular forms on general congruence subgroups, lattice-functional viewpoint, the Petersson inner product); Miyake *Modular Forms* (Springer 1989/2006) Ch. 2 (modular forms on $\mathrm{SL}_2(\mathbb{Z})$ and congruence subgroups); Koblitz *Introduction to Elliptic Curves and Modular Forms* (Springer GTM 97, 2nd ed. 1993) Ch. III (elementary exposition emphasising the bridge to elliptic curves)

Intuition [Beginner]

A modular form is a function on a particular two-dimensional region — the upper half-plane, the set of complex numbers with positive imaginary part — that exhibits a remarkable amount of symmetry. The symmetry comes from a discrete group of motions of the half-plane, the modular group, generated by the two transformations $τ \mapsto τ + 1$ (horizontal translation) and $τ \mapsto - 1/ τ$ (inversion through the unit semicircle). A function on the half-plane is modular of weight $k$ if, under every motion in the modular group, it scales by a precise power of the denominator of the motion. The weight $k$ records the power.

The intuition is that modular forms are functions admitting a special "symmetry book-keeping" property under a Möbius-symmetric tiling of the upper half-plane by a single fundamental polygon. Picture the upper half-plane carved into infinitely many copies of one curved triangle, with the modular group acting by gluing them together; a modular form is a function on a single triangle that knows how to extend to all the others by a fixed scaling rule. Adding the requirement that the function be holomorphic — analytic in the complex sense — and bounded near the cusp at infinity forces a striking finiteness: for each weight $k$ , there are only finitely many independent modular forms.

The deeper reason these functions matter is that they encode arithmetic information. The two basic examples, the Eisenstein series $E_{4}$ and $E_{6}$ , have Fourier expansions whose coefficients are divisor sums — pure number theory dropped into the analytic picture. The modular discriminant $Δ$ — the difference of cubes $E_{4}^{3}$ and squares $E_{6}^{2}$ normalised by $1728$ — has Fourier coefficients $τ (1), τ (2), τ (3), \dots$ that are the Ramanujan tau values, and these turn out to satisfy multiplicativity relations no one expected before Hecke proved them in $1937$ . Modular forms are the bridge between complex analysis and the arithmetic of integers; the bridge runs through the symmetry of the modular group and the rigidity of holomorphy.

Visual [Beginner]

A picture of the upper half-plane with the standard fundamental domain of the modular group drawn in: the region bounded by the two vertical lines at real part $- 1/2$ and $1/2$ and the unit semicircle below, forming a curved triangle with one vertex at the cusp at infinity, one vertex at the imaginary number $i$ , and one vertex at the cube root of unity $ρ = e^{2 π i /3}$ . Surrounding triangles are translates of this fundamental triangle under the action of the modular group, tiling the upper half-plane. A modular form takes a single value on each triangle that determines, by the weight- $k$ transformation rule, its values on every other triangle.

$The standard fundamental domain of the modular group $\mathrm{SL}_2(\mathbb{Z})$ in the upper half-plane: a curved triangle with vertices at the cusp $\infty$, at $i$, and at $\rho = e^{2 \pi i / 3}$, together with several neighbouring translates tiling the half-plane.$

The picture conveys the central feature: a modular form is determined by its values on one fundamental triangle plus a fixed scaling rule, and the resulting function on the upper half-plane is invariant in a precise, weight- $k$ -scaled sense under the entire tiling action.

Worked example [Beginner]

Pick the simplest case: a modular form of weight $4$ for the full modular group. The basic example is the Eisenstein series $E_{4}$ . We will compute its first few Fourier coefficients and verify the weight- $4$ transformation under one of the two generators of the modular group.

Step 1. The Fourier expansion of $E_{4}$ . With $q = e^{2 π i τ}$ , the normalised Eisenstein series of weight $4$ has the expansion $E_{4} (τ) = 1 + 240 \cdot σ_{3} (1) \cdot q + 240 \cdot σ_{3} (2) \cdot q^{2} + 240 \cdot σ_{3} (3) \cdot q^{3} + \dots,$ where $σ_{3} (n)$ is the sum of the cubes of the divisors of $n$ . Compute the first three: $σ_{3} (1) = 1$ , $σ_{3} (2) = 1 + 8 = 9$ , $σ_{3} (3) = 1 + 27 = 28$ . The series begins $E_{4} (τ) = 1 + 240 q + 2160 q^{2} + 6720 q^{3} + \dots$ .

Step 2. Check the first transformation rule. The translation $τ \mapsto τ + 1$ has denominator $c τ + d = 1$ (with $a = 1, b = 1, c = 0, d = 1$ ). The weight- $4$ rule demands $E_{4} (τ + 1) = (0 \cdot τ + 1)^{4} \cdot E_{4} (τ) = E_{4} (τ)$ . The Fourier expansion makes this evident: since $q = e^{2 π i τ}$ and $τ \mapsto τ + 1$ sends $q \mapsto e^{2 π i} q = q$ , the Fourier series is unchanged. Translation invariance is built in.

Step 3. Compute the value at the special point $τ = i \infty$ . As $τ$ moves up to infinity along the imaginary axis, $q = e^{2 π i τ} = e^{- 2 π \cdot Im (τ)} \to 0$ . So $E_{4} (i \infty) = 1$ . The Eisenstein series approaches the constant value $1$ at the cusp.

What this tells us: a modular form has both a numerical value at each finite point of the upper half-plane and a "value at infinity" given by the constant term of its Fourier expansion. The Eisenstein series $E_{4}$ takes the value $1$ at infinity; the discriminant $Δ$ vanishes there. The two examples already cover the two basic phenomena: holomorphic-but-nonzero at infinity (Eisenstein) and holomorphic-vanishing at infinity (cusp form).

Check your understanding [Beginner]

Exercise (easy, multiple choice).

A modular form of weight $k$ for the modular group $SL_{2} (Z)$ is a holomorphic function $f$ on the upper half-plane satisfying, under every matrix $(a c b d)$ in the modular group, the transformation rule:

A. $f ((a τ + b) / (c τ + d)) = f (τ)$ (plain invariance).
B. $f ((a τ + b) / (c τ + d)) = (c τ + d)^{k} f (τ)$ (weight- $k$ factor of automorphy).
C. $f ((a τ + b) / (c τ + d)) = (a τ + b)^{k} f (τ)$ (numerator power).
D. $f ((a τ + b) / (c τ + d)) = (det γ)^{k} f (τ)$ (determinant power, which is $1$ ).

Hint

The factor of automorphy involves the denominator of the Möbius transformation. The determinant of every matrix in $SL_{2} (Z)$ is $1$ , so a determinant power would be the constant function $1$ — not a useful rule.

Answer

B. The weight- $k$ transformation rule is $f ((a τ + b) / (c τ + d)) = (c τ + d)^{k} f (τ)$ for every $γ = (a c b d) \in SL_{2} (Z)$ . The factor $(c τ + d)$ is the automorphy factor of the Möbius transformation at the point $τ$ , and it equals the derivative of the Möbius map up to a sign; raising it to the $k$ -th power produces a rule consistent under composition.

Concretely, applying $γ_{1}$ then $γ_{2}$ multiplies $f$ by the product of the two automorphy factors, matching the automorphy factor of the composite $γ_{2} γ_{1}$ . Option A (plain invariance) describes a weight- $0$ modular function such as the $j$ -invariant. Option C (numerator power) is not consistent under composition. Option D collapses to the constant function $1$ because every element of $SL_{2} (Z)$ has determinant $1$ .

Formal definition [Intermediate+]

We work over the complex numbers. Let $H = {τ \in C : Im (τ) > 0}$ denote the open upper half-plane, and let $SL_{2} (Z) = {γ = (a c b d) : a, b, c, d \in Z, a d - b c = 1}$ denote the modular group of integer $2 \times 2$ matrices of determinant $1$ . The group $SL_{2} (Z)$ acts on $H$ by Möbius transformations $γ \cdot τ = \frac{a τ + b}{c τ + d}, γ = (a c b d),$ and the automorphy factor at $(γ, τ)$ is the complex number $j (γ, τ) := c τ + d$ .

Definition (modular form of weight $k$ ). Let $k$ be a non-negative integer. A modular form of weight $k$ for $SL_{2} (Z)$ is a holomorphic function $f : H \to C$ satisfying:

(i) Modular transformation law. For every $γ \in SL_{2} (Z)$ and every $τ \in H$ , $$ f(\gamma \cdot \tau) ;=; (c \tau + d)^k \cdot f(\tau). $$

(ii) Holomorphy at the cusp $\infty$ . The function $f$ is bounded as $Im (τ) \to + \infty$ . Equivalently, writing $q = e^{2 π i τ}$ and using the translation invariance $f (τ + 1) = f (τ)$ from (i) with $γ = T = (1011)$ , the function $f$ admits a Fourier expansion $$ f(\tau) ;=; \sum_{n \in \mathbb{Z}} a_n(f) , q^n, $$ and the condition (ii) is the vanishing $a_{n} (f) = 0$ for every $n < 0$ .

The complex vector space of all modular forms of weight $k$ for $SL_{2} (Z)$ is denoted $M_{k} (SL_{2} (Z))$ , or simply $M_{k}$ when the level is understood.

Definition (cusp form). A cusp form of weight $k$ is a modular form $f \in M_{k}$ satisfying additionally $a_{0} (f) = 0$ — equivalently, $f$ vanishes at the cusp $\infty$ . The complex vector space of cusp forms is denoted $S_{k} \subseteq M_{k}$ , where the letter $S$ derives from the German Spitzenform.

Definition (Eisenstein series). For every even integer $k \geq 4$ , the Eisenstein series of weight $k$ is the lattice sum $$ G_k(\tau) ;=; \sum_{\substack{(m, n) \in \mathbb{Z}^2 \ (m, n) \neq (0, 0)}} \frac{1}{(m \tau + n)^k}, \qquad \tau \in \mathbb{H}, $$ which converges absolutely and uniformly on compact subsets of $H$ for $k \geq 3$ . The normalised Eisenstein series is $$ E_k(\tau) ;=; \frac{G_k(\tau)}{2 \zeta(k)} ;=; 1 - \frac{2k}{B_k} \sum_{n \geq 1} \sigma_{k - 1}(n) , q^n, $$ where $B_{k}$ is the $k$ -th Bernoulli number and $σ_{k - 1} (n) = \sum_{d ∣ n} d^{k - 1}$ is the divisor power sum.

Definition (modular discriminant). The modular discriminant is the weight- $12$ cusp form $$ \Delta(\tau) ;=; \frac{E_4(\tau)^3 - E_6(\tau)^2}{1728} ;=; q \prod_{n \geq 1} (1 - q^n)^{24}, $$ the second equality the product formula due to Jacobi (and rederived by Hecke). The Fourier coefficients $τ (n)$ of $Δ$ are the Ramanujan tau function: $Δ (τ) = \sum_{n \geq 1} τ (n) q^{n}$ with $τ (1) = 1, τ (2) = - 24, τ (3) = 252, τ (4) = - 1472, \dots$ .

Counterexamples to common slips [Intermediate+]

The weight $k$ is a non-negative integer, not an arbitrary real number. For $SL_{2} (Z)$ , the matrix $- I = (- 1 0 0 - 1)$ is in the group, and the transformation law evaluates as $f (τ) = (- 1)^{k} f (τ)$ . For odd $k$ this forces $f \equiv 0$ , so $M_{k} (SL_{2} (Z)) = 0$ for every odd $k$ . The interesting cases are even weights.
The lattice sum in $G_{k}$ is over $(m, n) \neq = (0, 0)$ , not over coprime $(m, n)$ . The two sums differ by a factor of $ζ (k)$ : the coprime sum is $G_{k} / (2 ζ (k))$ once the $\pm (m, n)$ identification is accounted for. The two normalisations of the Eisenstein series in the literature, $G_{k}$ and $E_{k}$ , differ by the explicit constant $2 ζ (k)$ — be careful which one a given source uses.
The "cusp" of $SL_{2} (Z)$ is a single point $\infty$ , not a finite set. The orbit of the imaginary-axis-infinity point under $SL_{2} (Z)$ contains every rational number on the boundary $R \cup {\infty}$ of $H$ , but these all collapse to a single cusp because they form one $SL_{2} (Z)$ -orbit. For congruence subgroups $Γ \subseteq SL_{2} (Z)$ , multiple cusps appear and the definition of holomorphy must be checked at each one separately.
Holomorphy at the cusp is a Fourier-coefficient condition, not a continuity condition. The function $1/Δ$ is holomorphic on $H$ and satisfies the weight $- 12$ transformation rule, but its Fourier expansion $1/Δ (τ) = q^{- 1} + 24 + 324 q + \dots$ has a pole at $\infty$ : it is a meromorphic modular form, not a modular form. The definition of $M_{k}$ rules out such functions by requiring $a_{n} = 0$ for $n < 0$ .

Key theorem with proof [Intermediate+]

Theorem (Hecke 1937 / Serre VII §3 Theorem 4). The graded ring of modular forms for the full modular group is the polynomial ring on $E_{4}$ and $E_{6}$ : $$ M_*(\mathrm{SL}2(\mathbb{Z})) ;=; \bigoplus{k \geq 0} M_k(\mathrm{SL}_2(\mathbb{Z})) ;\cong; \mathbb{C}[E_4, E_6], $$ where $E_{4}$ has grading degree $4$ and $E_{6}$ has grading degree $6$ . The two generators $E_{4}$ and $E_{6}$ are algebraically independent over $C$ .

Proof. The proof is the residue-theorem argument on the fundamental domain, due to Hecke 1937 in the level- $1$ case and reproduced in Serre VII §3.

Step 1 (the valence formula). The standard fundamental domain of $SL_{2} (Z)$ acting on $H$ is $$ F ;=; \left{ \tau \in \mathbb{H} ,:, |\tau| \geq 1, ; |\mathrm{Re}(\tau)| \leq 1/2 \right}, $$ a curved triangle with vertices at the cusp $\infty$ , at $i$ , and at $ρ = e^{2 π i /3}$ . Let $f \in M_{k}$ be a non-zero modular form of weight $k$ , and let $ν_{P} (f)$ denote the order of vanishing of $f$ at a point $P \in F$ (with $ν_{\infty} (f) := min {n : a_{n} (f) \neq = 0}$ the cusp order, and $ν_{i} (f), ν_{ρ} (f)$ the half-integer-counted orders at the elliptic points). Then the valence formula holds: $$ \nu_\infty(f) ;+; \tfrac{1}{2} \nu_i(f) ;+; \tfrac{1}{3} \nu_\rho(f) ;+; \sum_{P \in F, P \neq i, \rho} \nu_P(f) ;=; \frac{k}{12}. $$ This is the residue theorem applied to $d lo g f = f^{'} / f d τ$ integrated around the boundary of $F$ . The boundary integral splits into four arcs: the two vertical sides cancel by translation periodicity (the same value of $f^{'} / f$ on opposite sides); the unit-semicircle arc is conjugated to itself by inversion $τ \mapsto - 1/ τ$ , contributing the half-integer terms at $i$ and $ρ$ ; the top arc near $\infty$ contributes the $- 2 π i ν_{\infty} (f)$ term. After computing the total integral as $- 2 π i \cdot k /12$ (the contour traversal picks up $k$ times $- 2 π i /12$ from the boundary winding), one extracts the valence formula in the displayed form.

Step 2 (algebraic independence of $E_{4}$ and $E_{6}$ ). Suppose a polynomial relation $P (E_{4}, E_{6}) = 0$ with $P \in C [X, Y]$ of total weighted degree $k$ (with $de g X = 4, de g Y = 6$ ). Each monomial $E_{4}^{a} E_{6}^{b}$ contributes a modular form of weight $4 a + 6 b$ . Grouping by weight, the relation becomes a sum $\sum_{k} P_{k} (E_{4}, E_{6}) = 0$ with each $P_{k}$ a modular form of weight $k$ , forcing each $P_{k} = 0$ separately. So assume $P$ is homogeneous of weight $k$ . Evaluate at the elliptic point $ρ$ : by the valence formula and the explicit factorisations $E_{4} (ρ) = 0$ and $E_{6} (ρ) \neq = 0$ (verified from the Fourier expansions plus the residue argument), each monomial $E_{4}^{a} E_{6}^{b}$ evaluates at $ρ$ to a non-zero multiple of $δ_{a, 0}$ times $E_{6} (ρ)^{b}$ . Distinct monomials give distinct $b$ values, so the relation forces all coefficients with $a = 0$ to vanish. Iterating with the elliptic point $i$ (where $E_{4} (i) \neq = 0, E_{6} (i) = 0$ ) eliminates the remaining coefficients. So $P = 0$ , proving algebraic independence.

Step 3 (generation by $E_{4}$ and $E_{6}$ ). Induct on weight $k$ . For $k < 4$ , the spaces $M_{0} = C$ and $M_{2} = 0$ (by the valence formula: $k /12 = 1/6$ is non-integer, so no non-zero form can satisfy it) and $M_{3} = 0$ (odd weight). For $k = 4$ : by the valence formula a non-zero form of weight $4$ has total vanishing-order sum $1/3$ ; the only way is $ν_{ρ} (f) = 1, ν_{\infty} = 0, ν_{i} = 0$ , and the constant-term normalisation pins $f = E_{4}$ up to scale, so $M_{4} = C \cdot E_{4}$ . Similarly $M_{6} = C \cdot E_{6}$ .

For $k \geq 8$ : given $f \in M_{k}$ , find a polynomial monomial $E_{4}^{a} E_{6}^{b}$ of the same weight ( $4 a + 6 b = k$ — solvable in non-negative integers because $k \geq 8$ and $k$ even). Subtract a scalar multiple to kill the constant term of $f$ : $f - c E_{4}^{a} E_{6}^{b} \in S_{k}$ is a cusp form of weight $k$ . Dividing by $Δ$ , a cusp form of weight $k$ becomes a modular form of weight $k - 12$ , and by induction this is a polynomial in $E_{4}, E_{6}$ . Multiplying back by $Δ = (E_{4}^{3} - E_{6}^{2}) /1728$ — itself a polynomial in $E_{4}, E_{6}$ — completes the induction.

Combining steps 2 and 3, every modular form is a polynomial in $E_{4}, E_{6}$ and the polynomial is unique. The structure theorem follows: $M_{*} (SL_{2} (Z)) ≅ C [E_{4}, E_{6}]$ as graded $C$ -algebras. $□$

Bridge. The Hecke / Serre structure theorem builds toward [21.04.02] Hecke operators and the Hecke algebra, where the $C [E_{4}, E_{6}]$ decomposition is refined by simultaneous eigenspace decomposition for the operators $T_{p}$ , and appears again in [21.04.03] Eichler-Shimura correspondence, where the weight- $2$ cusp forms — invisible at level $1$ but central at higher level $Γ_{0} (N)$ — pair with $2$ -dimensional $ℓ$ -adic Galois representations. The foundational reason is that the valence formula constrains modular forms so tightly that the entire graded ring is finitely generated by two elements; this is exactly the bridge between the analytic-symmetry definition and the arithmetic content carried by Fourier coefficients. Putting these together identifies the space $M_{k}$ of weight- $k$ modular forms with an explicit finite-dimensional vector space whose dimension is computable from $k$ alone, generalises to congruence subgroups via the modular-curve framework, and is dual to the lattice viewpoint expressed through Hecke's original definition of modular forms as functions on lattices in $C$ .

Exercises [Intermediate+]

A graded set covering the modular-group action, Eisenstein-series Fourier expansions, the valence formula, the structure theorem, and basic dimension counting.

Exercise 3 (medium, symbolic).

Show that $M_{2} (SL_{2} (Z)) = 0$ using the valence formula.

Hint

The valence formula gives $ν_{\infty} + \frac{1}{2} ν_{i} + \frac{1}{3} ν_{ρ} + \sum ν_{P} = k /12$ . For $k = 2$ , the right side is $1/6$ . The left side is a sum of non-negative rationals; check which combinations of half-integers and thirds can sum to $1/6$ .

Answer

Suppose $f \in M_{2} (SL_{2} (Z))$ is non-zero. The valence formula gives $$ \nu_\infty(f) + \tfrac{1}{2} \nu_i(f) + \tfrac{1}{3} \nu_\rho(f) + \sum_P \nu_P(f) = \tfrac{1}{6}. $$ The integer parts $ν_{\infty}, \sum_{P} ν_{P}$ and the half-integer / third-integer parts $ν_{i} /2, ν_{ρ} /3$ must combine to $1/6$ . Listing the possibilities: $ν_{ρ} = 1$ contributes $1/3 > 1/6$ , ruled out; $ν_{ρ} = 0$ leaves the equation $ν_{\infty} + \frac{1}{2} ν_{i} + \sum_{P} ν_{P} = 1/6$ , which requires $ν_{i} = 1/3$ , contradicting $ν_{i} \in Z_{\geq 0}$ . No non-negative-integer solution exists, so $f = 0$ and $M_{2} = 0$ .

Exercise 5 (medium, short-answer).

State the dimensions $dim M_{k} (SL_{2} (Z))$ for $k = 0, 2, 4, 6, 8, 10, 12, 14$ . Identify which are zero and which contain a one-dimensional space of cusp forms.

Hint

Use the dimension formula and recall $dim S_{k} = dim M_{k} - 1$ for $k \geq 4$ even (and zero otherwise), since the Eisenstein contribution is one-dimensional and complementary to the cusp-form subspace.

Answer

$k$	$dim M_{k}$	$dim S_{k}$	Comment
$0$	$1$	$0$	Constants.
$2$	$0$	$0$	Valence formula obstruction.
$4$	$1$	$0$	$M_{4} = C E_{4}$ .
$6$	$1$	$0$	$M_{6} = C E_{6}$ .
$8$	$1$	$0$	$M_{8} = C E_{4}^{2} = C E_{8}$ .
$10$	$1$	$0$	$M_{10} = C E_{4} E_{6} = C E_{10}$ .
$12$	$2$	$1$	$M_{12} = C E_{4}^{3} \oplus C Δ$ ; first cusp form appears at weight $12$ .
$14$	$1$	$0$	$M_{14} = C E_{4}^{2} E_{6} = C E_{14}$ .

The first cusp form is the modular discriminant $Δ \in S_{12}$ , which is the structural reason why every Hecke eigenform of small weight is forced to be Eisenstein and the modular forms below weight $12$ are entirely understood from $E_{4}, E_{6}$ alone.

Exercise 6 (hard, symbolic).

Derive the identity $E_{4}^{3} - E_{6}^{2} = 1728Δ$ by computing the first two non-zero Fourier coefficients of both sides.

Hint

Use $E_{4} = 1 + 240 q + \dots$ and $E_{6} = 1 - 504 q + \dots$ . Expand both cubes and squares through $q^{1}$ and $q^{2}$ ; subtract.

Answer

Through $q^{2}$ , $E_{4} = 1 + 240 q + 2160 q^{2}$ and $E_{6} = 1 - 504 q + α q^{2}$ where $α = - 16632$ (from $- 504 \cdot σ_{5} (2) = - 504 \cdot 33 = - 16632$ ).

Cube: $E_{4}^{3} = (1 + 240 q + 2160 q^{2} + \dots)^{3}$ . The coefficient of $q^{0}$ is $1$ , of $q^{1}$ is $3 \cdot 240 = 720$ , of $q^{2}$ is $3 \cdot 2160 + 3 \cdot 24 0^{2} = 6480 + 172800 = 179280$ .

Square: $E_{6}^{2} = (1 - 504 q + α q^{2} + \dots)^{2}$ . The coefficient of $q^{0}$ is $1$ , of $q^{1}$ is $- 2 \cdot 504 = - 1008$ , of $q^{2}$ is $2 α + 50 4^{2} = - 33264 + 254016 = 220752$ .

Subtract: $E_{4}^{3} - E_{6}^{2}$ has coefficients $[q^{0}] = 0$ , $[q^{1}] = 720 + 1008 = 1728$ , $[q^{2}] = 179280 - 220752 = - 41472$ . Dividing by $1728$ : $Δ = q + (- 41472/1728) q^{2} + \dots = q - 24 q^{2} + \dots$ , matching the known expansion $Δ = q - 24 q^{2} + 252 q^{3} - 1472 q^{4} + \dots$ . The identity $E_{4}^{3} - E_{6}^{2} = 1728Δ$ holds; the constant term vanishes because $Δ$ is a cusp form, and the normalisation factor $1728$ pins down the leading coefficient of $Δ$ to $1$ .

Exercise 7 (hard, short-answer).

Explain why the Eisenstein series $G_{2} (τ) = \sum_{(m, n) \neq = (0, 0)} (m τ + n)^{- 2}$ is not a modular form, and what its modular-transformation defect is.

Hint

The lattice sum for $k = 2$ is only conditionally convergent. The order of summation matters; symmetric summation produces a holomorphic function on $H$ but with an extra term in the modular transformation rule.

Answer

The double sum $\sum_{(m, n) \neq = (0, 0)} (m τ + n)^{- 2}$ fails to converge absolutely because the partial-sum analysis gives a logarithmic divergence at infinity. Defining $G_{2}$ by symmetric summation — fixing the outer sum on $m$ and then summing on $n$ — produces a holomorphic function on $H$ , but the resulting function satisfies the quasi-modular transformation rule $$ G_2(\gamma \cdot \tau) = (c \tau + d)^2 G_2(\tau) - 2 \pi i c (c \tau + d), \qquad \gamma = \begin{pmatrix} a & b \ c & d \end{pmatrix} \in \mathrm{SL}2(\mathbb{Z}). $$ The extra term $- 2 π i c (c τ + d)$ obstructs full modularity. The function $E_2(\tau) := G_2(\tau) / (2 \zeta(2)) = 1 - 24 \sum{n \geq 1} \sigma_1(n) q^n $i s t h er e f or e a * * q u a s i - m o d u l a r f or m * * o f w e i g h t$ 2 $r a t h er t hana t r u e m o d u l a r f or m . T h es p a ce$ M_2(\mathrm{SL}_2(\mathbb{Z})) $r e main sz er o . T h e f u n c t i o n$ E_2 $i s, h o w e v er, u se f u l : co mbina t i o n s l ik e$ E_2(\tau) - 2 E_2(2 \tau) $a r e m o d u l a r o f w e i g h t$ 2 $f or t h eco n g r u e n ces u b g r o u p$ \Gamma_0(2)$, and quasi-modular forms play a central role in the theory of mock modular forms and the topological-string-theory partition functions.

Exercise 8 (hard, short-answer).

Discuss the role of the Petersson inner product on $S_{k} (SL_{2} (Z))$ and the orthogonal decomposition $M_{k} = C \cdot E_{k} \oplus S_{k}$ .

Hint

The Petersson inner product is the $L^{2}$ -inner product on $S_{k}$ with respect to the hyperbolic measure on the fundamental domain. It is invariant under Hecke operators (self-adjoint), which gives the eigenform decomposition.

Answer

The Petersson inner product on the space of weight- $k$ cusp forms is $$ \langle f, g \rangle ;=; \int_F f(\tau) \overline{g(\tau)} , y^k , \frac{dx , dy}{y^2}, $$ with $F$ the standard fundamental domain, $τ = x + i y$ , and the hyperbolic measure $y^{- 2} d x d y$ ensuring $SL_{2} (Z)$ -invariance of the integrand. The integral converges absolutely on cusp forms because the cusp-vanishing condition forces $∣ f (τ) ∣ = O (e^{- 2 π y})$ as $y \to \infty$ . The inner product is positive-definite Hermitian and makes $S_{k}$ into a finite-dimensional Hilbert space.

Within $M_{k}$ for $k \geq 4$ even, the orthogonal decomposition with respect to the natural extension of the Petersson product is $$ M_k(\mathrm{SL}_2(\mathbb{Z})) ;=; \mathbb{C} \cdot E_k ;\oplus; S_k(\mathrm{SL}_2(\mathbb{Z})), $$ where $E_{k}$ is the normalised Eisenstein series of weight $k$ and the inner product is extended to allow Eisenstein contributions via regularised integration. The Eisenstein component is orthogonal to every cusp form by an explicit unfolding-integral argument (Rankin-Selberg unfolding). The Hecke operators $T_{n}$ , defined in [21.04.02], are self-adjoint with respect to $⟨ \cdot, \cdot ⟩$ ; the spectral theorem gives an orthogonal basis of $S_{k}$ by simultaneous Hecke eigenforms, the Hecke eigenform basis. Each eigenform $f$ has a Fourier expansion $f (τ) = \sum a_{n} q^{n}$ with multiplicative coefficients, and the attached $L$ -function $L (f, s) = \sum a_{n} / n^{s}$ has an Euler product and a functional equation, the prototype of the Langlands-functoriality program. The Petersson inner product is the analytic backbone making the Hecke-eigenform decomposition possible.

Lean formalisation [Intermediate+]

The Lean module Codex.NumberTheory.ModularForms.SL2Z schematises the data and key theorems of the modular-form theory for the full modular group. Current Mathlib supplies the underlying analytic infrastructure: Mathlib.Analysis.SpecialFunctions.Complex.Log and Mathlib.Analysis.Analytic.Basic for holomorphic functions, Mathlib.LinearAlgebra.Matrix.SpecialLinearGroup for the modular group $SL_{2} (Z) = SpecialLinearGroup (Fin 2) Z$ , and Mathlib.NumberTheory.Modular for the action on the upper half-plane.

The module declares typed placeholders for the seven essential structures of the theory: the upper half-plane UpperHalfPlane, the modular group ModularGroup, the Möbius action mobius and automorphy factor automorphyFactor, the structure ModularForm k recording the weight- $k$ transformation law and holomorphy at infinity, the substructure CuspForm k of forms vanishing at the cusp, the Eisenstein series eisensteinSeries k for even $k \geq 4$ , and the modular discriminant modularDiscriminant : CuspForm 12.

Three named theorems are recorded with sorry-equivalent proof bodies: modular_ring_polynomial_in_E4_E6 recording the Hecke / Serre structure theorem $M_{*} (SL_{2} (Z)) ≅ C [E_{4}, E_{6}]$ ; dim_formula_even and dim_formula_odd recording the dimension formula split on parity of $k$ modulo $12$ ; and discriminant_via_E4_E6 recording the explicit identity $1728Δ = E_{4}^{3} - E_{6}^{2}$ . The Mathlib gap enumerated in the frontmatter lean_mathlib_gap is the upstream-contribution roadmap for porting the foundational modular-form theory to Mathlib: lattice-sum convergence, holomorphy of Eisenstein series, the valence formula via the residue theorem on the fundamental domain, the dimension formula, and the Hecke-Serre polynomial-ring identification.

Modular forms, lattices, and the fundamental domain [Master]

The classical theory of modular forms admits two equivalent perspectives, both due to Hecke in his foundational papers: the upper-half-plane viewpoint in which $τ \in H$ parametrises lattices via $Λ_{τ} = Z τ + Z$ , and the lattice-functional viewpoint in which a modular form is a function on the set of lattices in $C$ satisfying a homogeneity relation. We articulate the equivalence and use it to derive the Eisenstein-series Fourier expansion.

Lattices and the upper half-plane. A lattice in $C$ is a discrete subgroup $Λ \subset C$ of rank $2$ , equivalently $Λ = Z ω_{1} + Z ω_{2}$ for two $R$ -linearly independent complex numbers $ω_{1}, ω_{2}$ . The ratio $τ = ω_{1} / ω_{2}$ lies in $C ∖ R$ ; replacing $Λ$ by $Λ/ ω_{2}$ scales the lattice to $Λ_{τ} = Z τ + Z$ , with $τ \in H$ after possibly swapping the order of the generators to ensure positive imaginary part. The set of lattices up to homothety $Λ \sim λ Λ$ (for $λ \in C^{*}$ ) is therefore in bijection with $SL_{2} (Z) \ H$ , the quotient of the upper half-plane by the modular group.

The Hecke lattice-functional definition. A function $F$ on the set of lattices in $C$ of weight $k$ satisfies $F (λ Λ) = λ^{- k} F (Λ)$ for every $λ \in C^{*}$ . Setting $f (τ) := F (Λ_{τ})$ produces a function on $H$ , and the lattice-homogeneity translates to the modular transformation law: for $γ \in SL_{2} (Z)$ , the lattice $Λ_{γ τ}$ is homothetic to $Λ_{τ}$ by the factor $(c τ + d)^{- 1}$ , so $f (γ τ) = (c τ + d)^{k} f (τ)$ . The lattice viewpoint is the perspective of Hecke 1937 and is the more natural framing for the link to elliptic curves: the lattice $Λ_{τ}$ is the period lattice of the elliptic curve $C / Λ_{τ}$ , and modular forms are sections of line bundles on the moduli space of elliptic curves $\overline{M}_{1, 1}$ .

Eisenstein-series Fourier expansion from the lattice viewpoint. The lattice-functional $G_{k} (Λ) = \sum_{ω \in Λ ∖ 0} ω^{- k}$ satisfies the weight- $k$ homogeneity and corresponds, via $G_{k} (τ) := G_{k} (Λ_{τ}) = \sum_{(m, n) \neq = (0, 0)} (m τ + n)^{- k}$ , to a modular form. To compute its Fourier expansion, split the sum on $m$ : the $m = 0$ contribution is $2 ζ (k)$ (twice because $\pm n$ both appear). For $m \neq = 0$ , group $\pm m$ pairs and use the Lipschitz formula

n \in Z \sum \frac{1}{( m τ + n ) ^{k}} = \frac{( - 2 π i ) ^{k}}{( k - 1 )!} d \geq 1 \sum d^{k - 1} q^{m d},

valid for $k \geq 2$ and $Im (τ) > 0$ , where $q = e^{2 π i τ}$ . Summing over $m \geq 1$ (and doubling for $\pm m$ ):

G_{k} (τ) = 2 ζ (k) + 2 \cdot \frac{( - 2 π i ) ^{k}}{( k - 1 )!} m \geq 1 \sum d \geq 1 \sum d^{k - 1} q^{m d} = 2 ζ (k) (1 - \frac{2 k}{B _{k}} n \geq 1 \sum σ_{k - 1} (n) q^{n}),

using the Euler identity $2 ζ (k) = - \frac{( 2 π i ) ^{k} B _{k}}{k !}$ for even $k$ to absorb the prefactor. Dividing by $2 ζ (k)$ yields the normalised expansion

E_{k} (τ) = 1 - \frac{2 k}{B _{k}} n \geq 1 \sum σ_{k - 1} (n) q^{n} .

For $k = 4$ : $B_{4} = - 1/30$ , so $- 2 k / B_{k} = - 8/ (- 1/30) = 240$ , giving the familiar $E_{4} = 1 + 240 \sum σ_{3} (n) q^{n}$ . For $k = 6$ : $B_{6} = 1/42$ , so $- 2 k / B_{k} = - 12/ (1/42) = - 504$ , giving $E_{6} = 1 - 504 \sum σ_{5} (n) q^{n}$ . The Eisenstein series are the source of an enormous family of arithmetic identities: every relation among modular forms of a given weight translates to a relation among divisor sums via the Fourier expansions.

The fundamental domain and the elliptic points. The standard fundamental domain $F = {τ \in H : ∣ τ ∣ \geq 1, ∣ Re (τ) ∣ \leq 1/2}$ has finite hyperbolic area $π /3$ — a remarkable finiteness given that $H$ itself has infinite area. The two elliptic points $i$ and $ρ = e^{2 π i /3}$ are the fixed points of the elliptic elements $S = (01 - 1 0)$ (order $2$ , fixing $i$ ) and $S T = (01 - 1 1)$ (order $3$ , fixing $ρ$ ). The compactification $X (1) = SL_{2} (Z) \ H \cup {\infty}$ is a Riemann surface of genus $0$ — the projective line — with three special points (two elliptic, one cusp). The valence formula reflects the order of stabilisers at the elliptic points: a modular form of weight $k$ on $X (1)$ has total divisor degree $k /12$ when half-counted at $i$ and third-counted at $ρ$ . The numerator $k$ is the form's weight; the denominator $12$ is the Euler characteristic of $X (1)$ with the orbifold structure $χ^{orb} (X (1)) = 1 - 1/2 - 2/3 = - 1/6$ inverted; the factor of $- 1$ reflects the canonical-bundle convention.

The graded ring $M_{*} = ⨁ M_{k}$ [Master]

The structure theorem $M_{*} (SL_{2} (Z)) ≅ C [E_{4}, E_{6}]$ identifies the entire graded ring of modular forms with an explicit polynomial ring on two generators. We unpack the consequences and refinements.

Cusp-form subring. The cusp forms $S_{*} = ⨁_{k} S_{k}$ form a graded ideal in $M_{*}$ , principally generated by $Δ$ : $$ S_*(\mathrm{SL}2(\mathbb{Z})) ;=; \Delta \cdot M{* - 12}(\mathrm{SL}2(\mathbb{Z})). $$ Equivalently, every cusp form is $Δ$ times a modular form of weight reduced by $12$ . The dimension formula $$ \dim S_k(\mathrm{SL}2(\mathbb{Z})) ;=; \dim M_k(\mathrm{SL}2(\mathbb{Z})) - 1 \quad \text{for $k \geq 4$ even} $$ follows: the Eisenstein contribution is one-dimensional, and the cusp-form complement has the dimension obtained by subtracting $1$ . The first non-zero cusp-form space is $S{12} = \mathbb{C} \cdot \Delta $, t h eseco n d i s$ S{16} = \mathbb{C} \cdot E_4 \Delta $, an d t h e d im e n s i o n s g r o w a ccor d in g t o$ \dim M{k - 12}$.

The Klein $j$ -invariant. The function $$ j(\tau) ;=; \frac{E_4(\tau)^3}{\Delta(\tau)} ;=; \frac{1}{q} + 744 + 196884 q + 21493760 q^2 + \cdots $$ is a modular function of weight $0$ — a meromorphic modular form with poles at the cusp. Its Fourier expansion is the moonshine series with coefficient $196884 = 196883 + 1$ matching the dimension of the smallest non-identity irreducible representation of the Monster sporadic finite simple group plus the dimension of the one-dimensional identity representation, the monstrous moonshine observation of McKay-Thompson (1979) proved by Borcherds (1992) ^{[Borcherds 1992]}. The $j$ -function generates the field of meromorphic modular functions on $X (1)$ : every modular function of weight $0$ is a rational function in $j$ . The map $j : X (1) \to P_{C}^{1}$ is a degree- $1$ map — an isomorphism of Riemann surfaces — sending $τ = i$ to $j = 1728$ , $τ = ρ$ to $j = 0$ , and $τ = \infty$ to $j = \infty$ .

Hilbert-Poincaré series. The dimension formula yields the Hilbert-Poincaré series of $M_{*} (SL_{2} (Z))$ as a graded $C$ -algebra: $$ H(M_*(\mathrm{SL}2(\mathbb{Z})), t) ;=; \sum{k \geq 0} \dim_\mathbb{C} M_k \cdot t^k ;=; \frac{1}{(1 - t^4)(1 - t^6)}. $$ The right-hand side is exactly the Hilbert-Poincaré series of the polynomial ring $C [E_{4}, E_{6}]$ with $de g E_{4} = 4, de g E_{6} = 6$ . The identity is the structure theorem in generating-function form: the equality of the Hilbert series determines the equality of the rings up to a graded isomorphism, and the algebraic independence of $E_{4}, E_{6}$ proven in the Key theorem ensures the isomorphism is induced by sending the abstract generators to the explicit Eisenstein series. Expanding $1/ ((1 - t^{4}) (1 - t^{6}))$ as a power series recovers the dimension formula entry-by-entry.

The ring $C [E_{4}, E_{6}]$ as a coordinate ring. Geometrically, $M_{*} (SL_{2} (Z))$ is the graded coordinate ring of the weighted projective space $P (4, 6)$ — the projective line $P^{1}$ with weights $4$ and $6$ on the two homogeneous coordinates. The morphism $P (4, 6) \to P^{1}$ given by $(E_{4} : E_{6}) \mapsto E_{4}^{3} / E_{6}^{2}$ presents $X (1)$ as a degree- $1$ cover of $P^{1}$ ramified over the three points ${0, 1728, \infty}$ , the values of the $j$ -invariant at the elliptic and cusp points. The line bundle $L_{k}$ on $X (1)$ whose sections are weight- $k$ modular forms is the $k$ -th power of a fundamental line bundle of orbifold-degree $1/12$ , and the dimension formula for $H^{0} (X (1), L_{k})$ is the orbifold Riemann-Roch theorem applied to $L_{k}$ .

Cusp forms and the Petersson inner product [Master]

The space $S_{k} (SL_{2} (Z))$ of cusp forms carries a canonical Hermitian inner product, the Petersson inner product, with respect to which the Hecke algebra acts self-adjointly. The Petersson product is the analytic infrastructure that makes the modular-form-to- $L$ -function dictionary work.

Definition of the Petersson inner product. Let $f, g \in S_{k} (SL_{2} (Z))$ be cusp forms of weight $k$ . The Petersson inner product is $$ \langle f, g \rangle ;=; \int_F f(\tau) \overline{g(\tau)} , (\mathrm{Im}(\tau))^k , \frac{dx , dy}{y^2}, $$ where $F$ is the standard fundamental domain, $τ = x + i y$ , and the hyperbolic measure $y^{- 2} d x d y$ ensures the integrand is $SL_{2} (Z)$ -invariant: under $γ = (a c b d)$ , the imaginary part transforms by $Im (γ τ) = Im (τ) /∣ c τ + d ∣^{2}$ , so $(Im (γ τ))^{k} = (Im (τ))^{k} /∣ c τ + d ∣^{2 k}$ , while $f (γ τ) \overline{g (γ τ)} = (c τ + d)^{k} \overline{(c τ + d)^{k}} f (τ) \overline{g (τ)} = ∣ c τ + d ∣^{2 k} f (τ) \overline{g (τ)}$ , and the two factors cancel. The hyperbolic measure $y^{- 2} d x d y$ is $SL_{2} (R)$ -invariant separately.

Convergence. The integral converges absolutely because $f$ and $g$ vanish at the cusp: as $y \to \infty$ , the Fourier expansion gives $∣ f (τ) ∣ = O (e^{- 2 π y})$ (the leading term is $a_{1} q$ with $∣ q ∣ = e^{- 2 π y}$ ), so the integrand decays exponentially in $y$ and the integral is finite. For modular forms that are not cusp forms, the integral diverges at the cusp and a regularised version (Rankin-Selberg unfolding) is required.

Hilbert-space structure. The Petersson inner product is positive-definite Hermitian: $⟨ f, f ⟩ > 0$ for $f \neq = 0$ , and $\overline{⟨ f, g ⟩} = ⟨ g, f ⟩$ . The space $S_{k}$ is finite-dimensional ( $dim M_{k} - 1$ for $k \geq 4$ even, otherwise zero), so the completion-to-a-Hilbert-space step is automatic: $S_{k}$ is already a finite-dimensional Hilbert space over $C$ . The orthogonal-decomposition theorems of finite-dimensional Hilbert spaces apply: any self-adjoint operator on $S_{k}$ admits an orthonormal basis of eigenvectors.

Self-adjointness of the Hecke operators. The Hecke operators $T_{n} : S_{k} \to S_{k}$ , defined in [21.04.02] via a sum over sublattices, are self-adjoint with respect to the Petersson inner product: $$ \langle T_n f, g \rangle ;=; \langle f, T_n g \rangle \qquad \text{for every } f, g \in S_k. $$ The proof unfolds the integral defining $T_{n}$ , transferring the lattice sum from one factor to the other; the symmetry is the central technical content of the Hecke theory. Self-adjointness has two consequences: (a) the eigenvalues of $T_{n}$ acting on $S_{k}$ are real; (b) the simultaneous Hecke eigenforms — forms that are simultaneously eigenfunctions of every $T_{n}$ — form an orthogonal basis of $S_{k}$ .

Hecke eigenform basis and Petersson normalisation. A Hecke eigenform is a non-zero $f \in S_{k}$ satisfying $T_{n} f = λ_{n} (f) f$ for every $n \geq 1$ . The eigenvalues $λ_{n} (f) \in R$ encode the multiplicative structure of the Fourier coefficients $a_{n} (f)$ : with the normalisation $a_{1} (f) = 1$ , the eigenform relation becomes $a_{n} (f) = λ_{n} (f)$ , i.e., the Hecke eigenvalues are exactly the Fourier coefficients. The Petersson product makes this normalisation canonical: every eigenform has a unique scaling with $a_{1} = 1$ , and distinct eigenforms (of the same weight) are Petersson-orthogonal because Hecke operators acting symmetrically have orthogonal eigenspaces.

The Petersson conjecture and Deligne's theorem. Petersson 1939 conjectured a sharp bound on the eigenvalues of $T_{p}$ acting on $S_{k}$ : for a weight- $k$ Hecke eigenform with $p$ -th Fourier coefficient $a_{p}$ , $$ |a_p| ;\leq; 2 p^{(k - 1)/2}. $$ This is the Ramanujan-Petersson conjecture, generalising Ramanujan's $1916$ conjecture that $∣ τ (p) ∣ \leq 2 p^{11/2}$ for the discriminant cusp form. The conjecture was proved by Deligne 1974 Publ. Math. IHES 43 ^{[Deligne 1974]} as a consequence of his proof of the Weil conjectures for the étale cohomology of Kuga-Sato varieties associated to modular curves. The Ramanujan-Petersson bound is the analytic strength built into the modular-form-to-Galois-representation dictionary of [21.04.03].

Synthesis. The Petersson inner product is the foundational reason that the analytic theory of cusp forms organises into a finite-dimensional Hilbert space carrying a self-adjoint Hecke algebra action, and the central insight is that the simultaneous-eigenform basis of $S_{k}$ identifies modular forms with Dirichlet series via the Mellin-transform correspondence $$ L(f, s) ;=; \sum_{n \geq 1} \frac{a_n(f)}{n^s} ;=; \prod_p \frac{1}{1 - a_p(f) p^{-s} + p^{k - 1 - 2 s}}, $$ the Euler product holding precisely when $f$ is a Hecke eigenform.

Putting these together identifies the analytic-symmetry definition of modular forms with the arithmetic-multiplicativity content carried by Hecke eigenvalues, this is exactly the bridge between automorphic forms and Galois representations developed in [21.04.03] and [21.05.01], and generalises to congruence subgroups $Γ_{0} (N), Γ_{1} (N)$ via the Atkin-Lehner new-form theory of the 1970s. The pattern recurs in the modularity theorem of Wiles-BCDT: every elliptic curve over $Q$ is the Mellin transform of a weight- $2$ Hecke eigenform on some $Γ_{0} (N)$ , and the dual to the Petersson product on cusp forms is the Mordell-Weil group of the elliptic curve. The Petersson inner product is the analytic spine of the entire arithmetic theory.

Connections [Master]

Riemann zeta function $ζ (s)$ 21.03.01. The Eisenstein series carry $ζ$ -values in their constant terms: $E_{k} = 1 - (2 k / B_{k}) \sum σ_{k - 1} (n) q^{n}$ and the leading coefficient $- 2 k / B_{k}$ is the same as $4 k / (2 ζ (1 - k))$ via the functional equation of $ζ$ . The $L$ -function of a Hecke eigenform $f$ is constructed as the Mellin transform of $f$ , mirroring the construction of $ζ (s)$ as the Mellin transform of a theta function; the modular-form / $L$ -function dictionary of Hecke 1936 generalises the Riemann-zeta picture from the constant automorphic form to weight- $k$ cusp forms.
Dirichlet $L$ -functions $L (s, χ)$ 21.03.02. Eisenstein series with character — variants $E_{k, χ}$ attached to a Dirichlet character $χ$ — have $L (s, χ)$ in their constant terms and provide the bridge between Dirichlet's analytic theory of primes in arithmetic progressions and the theory of modular forms on congruence subgroups. The Hecke operators in the modular-form setting are the multiplicative analogue of the Dirichlet-character twist.
Hecke operators and Hecke algebra 21.04.02. The Hecke algebra $T$ acts on $M_{k}$ commuting with the modular group action, and the simultaneous-eigenform basis of $S_{k}$ is the central technical refinement of the structure theorem of the present unit. Every Hecke eigenform $f$ has multiplicative Fourier coefficients $a_{n} (f)$ satisfying $a_{p} a_{q} = a_{pq}$ for distinct primes and the Hecke recursion $a_{p} a_{p^{r}} = a_{p^{r + 1}} + p^{k - 1} a_{p^{r - 1}}$ at prime powers; the structure theorem of the present unit identifies the ambient ring on which Hecke operators act, and [21.04.02] describes the operators themselves.
Eichler-Shimura correspondence 21.04.03. Weight- $2$ cusp eigenforms on $Γ_{0} (N)$ correspond to $2$ -dimensional $ℓ$ -adic Galois representations on the Tate module of the Jacobian of the modular curve $X_{0} (N)$ . The level- $1$ case treated in the present unit has no weight- $2$ cusp forms ( $M_{2} (SL_{2} (Z)) = 0$ by the valence formula), so the Eichler-Shimura correspondence carries substantive content only at higher levels — but the framework of [21.04.03] is built on the foundational theory developed here.
Möbius transformations 06.01.08. The modular group acts on the upper half-plane by Möbius transformations $τ \mapsto (a τ + b) / (c τ + d)$ , and the entire definition of modular forms is the requirement of a controlled transformation law under this action. The Möbius-transformation framework of [06.01.08] supplies the geometric substrate; the modular-form theory adds the arithmetic constraint that $γ \in SL_{2} (Z)$ rather than $γ \in SL_{2} (R)$ .
Holomorphic function 06.01.01. Modular forms are holomorphic functions on the upper half-plane satisfying additional symmetry and boundedness conditions; the entire theory rests on the rigidity of holomorphy, which forces a finite-dimensional dimension formula from a constraint that, for arbitrary continuous functions, would leave an infinite-dimensional space. The maximum modulus principle and the residue theorem of [06.01.01] are the analytic tools the valence-formula proof rests on.
Riemann surfaces and the modular curve $X (1)$ 06.03.01. The quotient $X (1) = SL_{2} (Z) \ H^{*}$ — the upper half-plane plus the cusp $\infty$ , modulo the modular group — is a compact Riemann surface of genus $0$ , isomorphic to $P^{1}$ via the $j$ -invariant. Modular forms are sections of line bundles on $X (1)$ , and the dimension formula for $M_{k}$ is the Riemann-Roch theorem for the bundle $L_{k}$ in the orbifold setting accounting for the two elliptic points $i$ and $ρ$ .
Linear transformation and rank-nullity 01.01.05. The matrix structure $γ = (a c b d) \in SL_{2} (Z)$ is a special linear transformation of $R^{2}$ with integer entries and determinant $1$ . The two generators $S = (01 - 1 0)$ and $T = (1011)$ of the modular group are the fundamental linear-transformation prerequisites for the Möbius action; the entire arithmetic content of the modular group emerges from the relation $(S T)^{3} = S^{2} = - I$ in the matrix group.
Dedekind / Hecke / Artin $L$ -functions 21.03.03. Successor unit on the higher-rank $L$ -function chapter. The $L$ -function of a Hecke eigenform $f \in S_{k} (SL_{2} (Z))$ is the $GL_{2} / Q$ automorphic shadow of an Artin $L$ -function: $L (s, f) = L (s, ρ_{f})$ for a $2$ -dimensional Galois representation $ρ_{f}$ , and Hecke 1936's modular-form / $L$ -function dictionary is the prototype $GL_{2}$ -side of the Hecke-Artin formalism developed in 21.03.03. The Mellin-transform construction of $L (s, f)$ mirrors the Hecke 1918-20 construction of Hecke $L$ -functions on number fields.
Modularity theorem and BSD 21.06.01. Successor unit on the elliptic-curve / modular-form bridge. The modularity theorem identifies every elliptic curve $E / Q$ with a weight- $2$ cusp newform $f_{E}$ on $Γ_{0} (N_{E})$ . The level- $1$ space $S_{2} (SL_{2} (Z)) = 0$ developed in the present unit is the boundary case (no weight- $2$ cusp forms at level $1$ , hence no level- $1$ modular elliptic curves); the modularity programme rests on the higher-level Atkin-Lehner extension of the foundational dimension-formula and $q$ -expansion machinery developed here.
Iwasawa $Z_{p}$ -extensions and $p$ -adic L-functions 21.07.01, 21.07.02. Successor units on the Iwasawa theory of modular forms. Hida's 1985 $Λ$ -adic modular forms organise weight- $k$ ordinary newforms into a single $Λ$ -adic family deforming over the Iwasawa algebra $Λ = Z_{p} [[T]]$ , with the weight-variable specialisation at each $k \geq 2$ producing a classical Hecke eigenform on $Γ_{0} (N p^{r})$ . The structure theory of $M_{k} (SL_{2} (Z))$ developed in the present unit is the fibrewise component of this Iwasawa-theoretic family, and the Mazur-Wiles 1984 proof of the Main Conjecture for $Q$ uses $Λ$ -adic Eisenstein series as the central technical input.

Historical & philosophical context [Master]

Modular forms were introduced by Eisenstein in the 1840s in his study of the lattice sums $\sum (m + n τ)^{- k}$ that now bear his name; the systematic theory crystallised with Klein and Fricke's late-nineteenth-century work on modular functions and the $j$ -invariant, particularly the Klein-Fricke 1890-92 Vorlesungen über die Theorie der elliptischen Modulfunktionen which collected the geometric theory of $SL_{2} (Z) \ H^{*}$ as a Riemann surface. Klein's 1879 Math. Ann. 14 paper on transformations of degree $7$ — the icosahedral connection — introduced the modern viewpoint that modular functions are coordinates on the moduli space of elliptic curves.

The arithmetic theory of modular forms was largely the achievement of Erich Hecke in the 1930s. Hecke's two 1936-37 papers in Mathematische Annalen (vol. 112 + 114) ^{[Hecke 1936; ref: TODO_REF Hecke 1937]} introduced the operators $T_{n}$ acting on $M_{k}$ , proved the multiplicativity of the Fourier coefficients of simultaneous eigenforms, derived the Euler product factorisation of the attached Dirichlet $L$ -function, and established the Hecke converse theorem identifying modular forms with Dirichlet series satisfying a functional equation. The structure theorem $M_{*} (SL_{2} (Z)) = C [E_{4}, E_{6}]$ recorded as the Key theorem of this unit appears in Hecke's 1937 paper II §2 in the level- $1$ case. Petersson 1939 ^{[Petersson 1939]} introduced the inner product bearing his name and the Ramanujan-Petersson conjecture on the size of Hecke eigenvalues.

The post-Hecke development extended the theory to congruence subgroups $Γ_{0} (N), Γ_{1} (N)$ and to weight- $1/2$ and higher-half-integer weights. Shimura's 1971 monograph Introduction to the Arithmetic Theory of Automorphic Functions synthesised the classical theory with the emerging arithmetic-geometry framework, and Atkin-Lehner 1970 Math. Ann. 185 introduced the new-form theory that gave a canonical-eigenform basis on $Γ_{0} (N)$ stable under twists by Dirichlet characters. Eichler 1954 and Shimura 1958 independently established the correspondence between weight- $2$ cusp eigenforms and $2$ -dimensional $ℓ$ -adic Galois representations — the bridge that Wiles 1995 and Breuil-Conrad-Diamond-Taylor 2001 would later use to prove the modularity theorem.

The Langlands programme, initiated by Langlands 1967-70, recast the theory of modular forms inside the framework of automorphic representations of $GL_{2} (A_{Q})$ , where $A_{Q}$ is the adele ring of $Q$ . The classical modular form on $SL_{2} (Z) \ H$ becomes the right $K_{\infty} \cdot K_{f}$ -fixed vector in an automorphic representation $π = π_{\infty} \otimes π_{f}$ , with $π_{\infty}$ a holomorphic discrete-series representation of $GL_{2} (R)$ and $π_{f}$ a representation of $GL_{2} (A_{f})$ . The Langlands functoriality conjectures connect $L$ -functions of modular forms to $L$ -functions of arbitrary automorphic representations on reductive groups, with the modular-form case the prototype for the entire programme. Borcherds 1992 ^{[Borcherds 1992]} proved the monstrous moonshine conjecture connecting the $j$ -invariant's Fourier coefficients to the representation theory of the Monster sporadic finite simple group, demonstrating that the modular-form framework extends to vertex operator algebras and string-theoretic partition functions.

Bibliography [Master]

@book{DiamondShurman2005,
  author = {Diamond, Fred and Shurman, Jerry},
  title = {A First Course in Modular Forms},
  series = {Graduate Texts in Mathematics},
  volume = {228},
  publisher = {Springer},
  year = {2005},
}

@book{Serre1973,
  author = {Serre, Jean-Pierre},
  title = {A Course in Arithmetic},
  series = {Graduate Texts in Mathematics},
  volume = {7},
  publisher = {Springer},
  year = {1973},
}

@book{Lang1976,
  author = {Lang, Serge},
  title = {Introduction to Modular Forms},
  series = {Grundlehren der mathematischen Wissenschaften},
  volume = {222},
  publisher = {Springer},
  year = {1976},
}

@book{Miyake1989,
  author = {Miyake, Toshitsune},
  title = {Modular Forms},
  series = {Springer Monographs in Mathematics},
  publisher = {Springer},
  year = {1989},
}

@book{Koblitz1993,
  author = {Koblitz, Neal},
  title = {Introduction to Elliptic Curves and Modular Forms},
  series = {Graduate Texts in Mathematics},
  volume = {97},
  edition = {2nd},
  publisher = {Springer},
  year = {1993},
}

@book{Shimura1971,
  author = {Shimura, Goro},
  title = {Introduction to the Arithmetic Theory of Automorphic Functions},
  series = {Publications of the Mathematical Society of Japan},
  volume = {11},
  publisher = {Princeton University Press / Iwanami Shoten},
  year = {1971},
}

@book{ManinPanchishkin2005,
  author = {Manin, Yuri I. and Panchishkin, Alexei A.},
  title = {Introduction to Modern Number Theory: Fundamental Problems, Ideas and Theories},
  series = {Encyclopaedia of Mathematical Sciences},
  volume = {49},
  edition = {2nd},
  publisher = {Springer},
  year = {2005},
}

@article{Hecke1936,
  author = {Hecke, Erich},
  title = {{Ü}ber die {B}estimmung {D}irichletscher {R}eihen durch ihre {F}unktionalgleichung},
  journal = {Mathematische Annalen},
  volume = {112},
  year = {1936},
  pages = {664--699},
}

@article{Hecke1937,
  author = {Hecke, Erich},
  title = {{Ü}ber {M}odulfunktionen und die {D}irichletschen {R}eihen mit {E}ulerscher {P}roduktentwicklung. {I}, {II}},
  journal = {Mathematische Annalen},
  volume = {114},
  year = {1937},
  pages = {1--28, 316--351},
}

@article{Petersson1939,
  author = {Petersson, Hans},
  title = {{Ü}ber die {E}ntwicklungskoeffizienten der automorphen {F}ormen},
  journal = {Acta Mathematica},
  volume = {58},
  year = {1932},
  pages = {169--215},
}

@article{Deligne1974,
  author = {Deligne, Pierre},
  title = {La conjecture de {W}eil. {I}},
  journal = {Publications Math\'ematiques de l'IH\'ES},
  volume = {43},
  year = {1974},
  pages = {273--307},
}

@article{Borcherds1992,
  author = {Borcherds, Richard E.},
  title = {Monstrous moonshine and monstrous {L}ie superalgebras},
  journal = {Inventiones Mathematicae},
  volume = {109},
  year = {1992},
  pages = {405--444},
}

@article{AtkinLehner1970,
  author = {Atkin, A. O. L. and Lehner, J.},
  title = {Hecke operators on $\Gamma_0(m)$},
  journal = {Mathematische Annalen},
  volume = {185},
  year = {1970},
  pages = {134--160},
}

@book{Bump1997,
  author = {Bump, Daniel},
  title = {Automorphic Forms and Representations},
  series = {Cambridge Studies in Advanced Mathematics},
  volume = {55},
  publisher = {Cambridge University Press},
  year = {1997},
}

Prerequisites

01.01.05
06.01.01
06.01.08

Tier anchors

beginner: Diamond-Shurman *A First Course in Modular Forms* (Springer GTM 228, 2005) §1.1 (informal motivating discussion of weight-$k$ transformation law and the modular group); Serre *A Course in Arithmetic* (Springer GTM 7, 1973) Ch. VII §1 introductory paragraphs
intermediate: Diamond-Shurman *A First Course in Modular Forms* §§1.1-1.2 (definition, Eisenstein series, $q$-expansions); Serre *A Course in Arithmetic* Ch. VII §§2-3 (Eisenstein series, the modular discriminant, the structure theorem)
master: Diamond-Shurman *A First Course in Modular Forms* §§1.1-1.5 + §§3.5-3.7 (full theory including dimension formula, valence formula, $L$-function attached to a cusp form); Serre *A Course in Arithmetic* Ch. VII §§1-4 (Hecke-style proof of $M_*(\mathrm{SL}_2(\mathbb{Z})) = \mathbb{C}[E_4, E_6]$ via the valence formula); Manin-Panchishkin *Introduction to Modern Number Theory* (Springer EMS 49, 2nd ed. 2005) Ch. 6 §§1-3 (modular forms in the encyclopaedic survey context, Hecke's lattice viewpoint, the bridge to $L$-functions); Hecke 1936-37 *Math. Ann.* 112 + 114 (originator papers: definition of modular forms with multiplier system, the Hecke correspondence between modular forms and Dirichlet series, the ring structure of $M_*(\mathrm{SL}_2(\mathbb{Z}))$); Shimura *Introduction to the Arithmetic Theory of Automorphic Functions* (Princeton Iwanami 1971) §§2.1-2.7 (modular forms on general congruence subgroups, lattice-functional viewpoint, the Petersson inner product); Miyake *Modular Forms* (Springer 1989/2006) Ch. 2 (modular forms on $\mathrm{SL}_2(\mathbb{Z})$ and congruence subgroups); Koblitz *Introduction to Elliptic Curves and Modular Forms* (Springer GTM 97, 2nd ed. 1993) Ch. III (elementary exposition emphasising the bridge to elliptic curves)

References

TODO_REF
Diamond, F. & Shurman, J. — A First Course in Modular Forms · Springer Graduate Texts in Mathematics 228 (2005). Ch. 1 (definition of modular forms, weight-$k$ transformation law, Eisenstein series $E_k$, the modular discriminant $\Delta$, the structure theorem $M_*(\mathrm{SL}_2(\mathbb{Z})) = \mathbb{C}[E_4, E_6]$ via the valence formula); Ch. 3 (the Petersson inner product and orthogonality of cusp forms to Eisenstein series); the standard modern textbook reference for the foundational theory.
TODO_REF
Serre, J.-P. — A Course in Arithmetic · Springer Graduate Texts in Mathematics 7 (1973). Ch. VII §§1-4 (modular forms on $\mathrm{SL}_2(\mathbb{Z})$: fundamental domain, definition, Eisenstein series, the valence formula via the residue theorem on the fundamental domain, the structure theorem $M_*(\mathrm{SL}_2(\mathbb{Z})) = \mathbb{C}[E_4, E_6]$, dimension formula); the compact rigorous exposition that became the standard graduate-textbook account before Diamond-Shurman.
TODO_REF
Manin, Y. I. & Panchishkin, A. A. — Introduction to Modern Number Theory · Springer Encyclopaedia of Mathematical Sciences 49 (2nd ed. 2005), Number Theory I. Ch. 6 §1 (modular forms as automorphic forms on the upper half-plane); §2 ($L$-functions of modular forms, the Mellin transform); §3 (Hecke operators and eigenforms, the link to the Eichler-Shimura programme); the encyclopaedic-survey framing positioning modular forms inside the broader landscape of modern arithmetic geometry.
TODO_REF
Hecke, E. — Über die Bestimmung Dirichletscher Reihen durch ihre Funktionalgleichung · *Mathematische Annalen* 112 (1936), 664-699. The originator paper establishing the correspondence between modular forms with multiplier system and Dirichlet series satisfying a functional equation; introduces the Hecke converse theorem foundational to the modular-form / $L$-function dictionary.
TODO_REF
Hecke, E. — Über Modulfunktionen und die Dirichletschen Reihen mit Eulerscher Produktentwicklung I, II · *Mathematische Annalen* 114 (1937), 1-28 + 316-351. The two-part originator paper introducing Hecke operators $T_n$ on the space of modular forms, the multiplicativity relation, the simultaneous-eigenform basis, and the Euler product factorisation of the attached $L$-function; foundational for `21.04.02`.
TODO_REF
Shimura, G. — Introduction to the Arithmetic Theory of Automorphic Functions · Princeton University Press / Iwanami Shoten (1971), Publications of the Mathematical Society of Japan 11. §§2.1-2.7 (modular forms on $\mathrm{SL}_2(\mathbb{Z})$ and on congruence subgroups $\Gamma_0(N), \Gamma_1(N)$, the lattice-functional viewpoint, the Petersson inner product, dimension formulas); the canonical advanced reference treating arithmetic and analytic aspects together.
TODO_REF
Miyake, T. — Modular Forms · Springer Monographs in Mathematics (translation of the 1976 Japanese original; English 1989, corrected 2nd printing 2006). Ch. 2 (modular forms on $\mathrm{SL}_2(\mathbb{Z})$, Eisenstein series, the fundamental domain, the valence formula and dimension formula); Ch. 4-6 (Hecke operators, new-form theory, Atkin-Lehner). The most comprehensive modern monograph.
TODO_REF
Koblitz, N. — Introduction to Elliptic Curves and Modular Forms · Springer Graduate Texts in Mathematics 97 (1984; 2nd ed. 1993). Ch. III (modular forms on $\mathrm{SL}_2(\mathbb{Z})$: definition, Eisenstein series, modular discriminant, ring structure, dimension formula); emphasis on the bridge from modular forms to elliptic curves via the $j$-invariant, and the eventual link to BSD and Wiles's modularity theorem.
TODO_REF
Lang, S. — Introduction to Modular Forms · Springer Grundlehren der mathematischen Wissenschaften 222 (1976). Ch. III (definition and basic properties of modular forms on the full modular group, Eisenstein series); Ch. X (cusp forms, Petersson inner product, the modular discriminant); the second classical graduate-textbook reference alongside Serre and Shimura.
TODO_REF
Bump, D. — Automorphic Forms and Representations · Cambridge Studies in Advanced Mathematics 55 (1997). Ch. 1 (modular forms on $\mathrm{SL}_2(\mathbb{Z})$ as the prototype of the automorphic-form / Langlands-programme framework); positions the classical theory inside its representation-theoretic generalisation.

Lean module

Codex.NumberTheory.ModularForms.SL2Z

Reviewer

TBD

Estimated time

beginner: 20m
intermediate: 50m
master: 90m

Intuition [Beginner]

Visual [Beginner]

Worked example [Beginner]

Check your understanding [Beginner]

Formal definition [Intermediate+]

Counterexamples to common slips [Intermediate+]

Key theorem with proof [Intermediate+]

Exercises [Intermediate+]

Lean formalisation [Intermediate+]

Modular forms, lattices, and the fundamental domain [Master]

The graded ring M∗​=⨁Mk​ [Master]

Cusp forms and the Petersson inner product [Master]

Connections [Master]

Historical & philosophical context [Master]

Bibliography [Master]

The graded ring $M_{*} = ⨁ M_{k}$ [Master]