riemann

The Lean code claims a proof of the Riemann Hypothesis. This site has not checked the claim yet. See what is checked.

Class field theory

Additive and Hecke characters, quadratic symbols, local and global class field theory, the Hasse principle for norms and Hermitian forms, and Tate's thesis with Hecke L-functions and epsilon factors.

Prerequisites: Number fields.

Class field theory describes the abelian extensions of a local or global field by the multiplicative group of the field. Its Hecke characters are the coefficients of the theta lifts and of Row0b, the local Artin map is a law of IsRecT, and Tate's thesis gives the L-functions and epsilon factors in ThetaIntegralsAgree. Chapter 4 defines places, completions, adeles and ideles.

Additive characters

A character of a locally compact abelian group GG is a continuous homomorphism G→ℂ×G \to \mathbb{C}^\times, unitary if its values have absolute value 11. The unitary characters form the Pontryagin dual G^\widehat G.

Definition (additive character). An additive character of a local field FvF_v is a nontrivial unitary character ψv\psi_v of the additive group FvF_v. An additive character of 𝔸F\mathbb{A}_F is a nontrivial unitary character ψ\psi of 𝔸F\mathbb{A}_F with ψ(F)=1\psi(F) = 1.

Example (the standard character). Put ψ∞(x)=e2πix\psi_\infty(x) = e^{2\pi i x} on ℝ\mathbb{R}, and ψp(x)=e−2πi{x}p\psi_p(x) = e^{-2\pi i \{x\}_p} on ℚp\mathbb{Q}_p, where {x}p∈ℤ[1/p]\{x\}_p \in \mathbb{Z}[1/p] is the fractional part, with x−{x}p∈ℤpx - \{x\}_p \in \mathbb{Z}_p. Then ψℚ=∏vψv\psi_{\mathbb{Q}} = \prod_v \psi_v is trivial on ℚ\mathbb{Q}. For a number field FF, the standard additive character of 𝔸F\mathbb{A}_F is ψℚ∘TrF/ℚ\psi_{\mathbb{Q}} \circ \mathrm{Tr}_{F/\mathbb{Q}}.

Proposition (self-duality, Tate, 1950). Fix an additive character ψv\psi_v of FvF_v. Each unitary character of FvF_v is x↦ψv(ax)x \mapsto \psi_v(ax) for exactly one a∈Fva \in F_v. Likewise each unitary character of 𝔸F\mathbb{A}_F trivial on FF is x↦ψ(ax)x \mapsto \psi(ax) for exactly one a∈Fa \in F.

At a finite place vv, ψv\psi_v is unramified if it is trivial on 𝒪v\mathcal{O}_v but not on ϖv−1𝒪v\varpi_v^{-1}\mathcal{O}_v. The standard character is unramified at each finite place where F/ℚF/\mathbb{Q} is unramified.

In TLiftK, Base holds an additive character ψ\psi of 𝔸F\mathbb{A}_F and std asks that it is standard; the good triples of LocalMatching ask that it is unramified at the inert places.

Hecke characters

Let KK be a number field.

Definition. A Hecke character of KK is a continuous homomorphism χ:𝔸K×→ℂ×\chi : \mathbb{A}_K^\times \to \mathbb{C}^\times trivial on K×K^\times, that is, a character of CKC_K. It is unitary if ∣χ∣=1|\chi| = 1. Each Hecke character is χ0∣⋅∣𝔸t\chi_0 |\cdot|_{\mathbb{A}}^{t} with χ0\chi_0 unitary and tt real.

Each Hecke character is a product χ=∏vχv\chi = \prod_v \chi_v of characters of Kv×K_v^\times, and continuity forces χv(𝒪v×)=1\chi_v(\mathcal{O}_v^\times) = 1 for almost all finite vv.

Definition (conductor). Let vv be finite and χv\chi_v a character of Kv×K_v^\times. Put Uv(0)=𝒪v×U_v^{(0)} = \mathcal{O}_v^\times and Uv(c)=1+ϖvc𝒪vU_v^{(c)} = 1 + \varpi_v^c \mathcal{O}_v for c≥1c \geq 1. The conductor exponent c(χv)c(\chi_v) is the least c≥0c \geq 0 with χv(Uv(c))=1\chi_v(U_v^{(c)}) = 1. The character χv\chi_v is unramified if c(χv)=0c(\chi_v) = 0 and ramified otherwise; if it is unramified, χv(ϖv)\chi_v(\varpi_v) does not depend on the uniformizer. The conductor of χ\chi is the ideal 𝔣χ=∏v∤∞𝔭vc(χv)\mathfrak f_\chi = \prod_{v \nmid \infty} \mathfrak p_v^{c(\chi_v)}.

An unramified χv\chi_v is x↦zv(x)x \mapsto z^{v(x)} with z=χv(ϖv)z = \chi_v(\varpi_v). So a Hecke character unramified outside a finite set SS gives a character of the fractional ideals prime to SS, with χ(𝔭v)=χv(ϖv)\chi(\mathfrak p_v) = \chi_v(\varpi_v).

Example. Since 𝔸ℚ×=ℚ×⋅(ℝ>0×∏pℤp×)\mathbb{A}_\mathbb{Q}^\times = \mathbb{Q}^\times \cdot \bigl(\mathbb{R}_{>0} \times \prod_p \mathbb{Z}_p^\times\bigr), a finite order Hecke character of ℚ\mathbb{Q} is a character of ℤ^×=lim←⁡(ℤ/Nℤ)×\widehat{\mathbb{Z}}^\times = \varprojlim (\mathbb{Z}/N\mathbb{Z})^\times, that is, a primitive Dirichlet character (Chapter 6). In one normalization, χ\chi of conductor NN has χ(𝔭p)=χ(p)\chi(\mathfrak p_p) = \chi(p) for p∤Np \nmid N; others use χ(p)−1\chi(p)^{-1}.

Infinity type and algebraic characters

At a real place, a character of Kv×=ℝ×K_v^\times = \mathbb{R}^\times is x↦sgn(x)ϵ∣x∣sx \mapsto \mathrm{sgn}(x)^{\epsilon} |x|^{s} with ϵ∈{0,1}\epsilon \in \{0, 1\} and s∈ℂs \in \mathbb{C}. At a complex place, a character of ℂ×\mathbb{C}^\times is z↦zpzˉq:=(z/∣z∣)p−q∣z∣p+qz \mapsto z^{p} \bar z^{q} := (z/|z|)^{p-q} |z|^{p+q} with p,q∈ℂp, q \in \mathbb{C} and p−q∈ℤp - q \in \mathbb{Z}. The family of these exponents over the infinite places is the infinity type of χ\chi.

Definition (algebraic Hecke character). A Hecke character χ\chi of KK is algebraic (of type A0A_0) if there are integers nσn_\sigma, one for each embedding σ:K→ℂ\sigma : K \to \mathbb{C}, with χ∞(x)=∏σσ(x)nσ\chi_\infty(x) = \prod_\sigma \sigma(x)^{n_\sigma} on the identity component of K∞×K_\infty^\times. For a CM field EE with CM type Φ\Phi, a unitary character has infinity type n=(nσ)σ∈Φn = (n_\sigma)_{\sigma \in \Phi} relative to Φ\Phi if χv(σ−1(z))=(z/∣z∣)nσ\chi_v(\sigma^{-1}(z)) = (z/|z|)^{n_\sigma} at the complex place vv of each σ∈Φ\sigma \in \Phi. Sign conventions differ between texts.

Proposition (Weil, 1955). Let χ\chi be an algebraic Hecke character of KK, ℓ\ell a prime and ι:ℚ‾ℓ≅ℂ\iota : \overline{\mathbb{Q}}_\ell \cong \mathbb{C} a field isomorphism. For one fixed normalization of Frobenius, there is a unique continuous character χℓ:Gal(K‾/K)→ℚ‾ℓ×\chi_\ell : \mathrm{Gal}(\overline{K}/K) \to \overline{\mathbb{Q}}_\ell^\times, the ℓ\ell-adic form of χ\chi, with ι χℓ(Frobv)=χ(ϖv)\iota\,\chi_\ell(\mathrm{Frob}_v) = \chi(\varpi_v) at each finite v∤ℓv \nmid \ell where χ\chi is unramified. The construction uses the global Artin map (defined below).

Example. For K=ℚK = \mathbb{Q}, the idele norm ∣⋅∣𝔸|\cdot|_{\mathbb{A}} is algebraic with n=1n = 1, being x↦xx \mapsto x on ℝ>0\mathbb{R}_{>0}. Its ℓ\ell-adic form is the ℓ\ell-adic cyclotomic character or its inverse, according to the normalization of Frobenius.

Row0b takes a Hecke character χ\chi of E+E^+ with χv(−1)=(−1)a\chi_v(-1) = (-1)^a at each infinite vv (sign). The good triples of the theta matching chapter use Hecke characters α,β\alpha, \beta of EE with infinity types mα,mβm_\alpha, m_\beta, unramified outside a finite set TT, with ν=α/β\nu = \alpha/\beta ramified at each place of TT. The ℓ\ell-adic form appears as heckeLAdicCharacter and heckeLAdicCharacterAbove in UnitaryGroupThirdPlaceHypotheses.

Quadratic symbols

Definition. Let pp be an odd prime and a∈ℤa \in \mathbb{Z}. The Legendre symbol (ap)\left(\frac{a}{p}\right) is 00 if p∣ap \mid a, 11 if aa is a nonzero square modulo pp, and −1-1 otherwise. For an odd positive integer n=p1⋯pkn = p_1 \cdots p_k (primes repeated), the Jacobi symbol is (an)=∏i(api)\left(\frac{a}{n}\right) = \prod_{i} \left(\frac{a}{p_i}\right).

For fixed nn, a↦(an)a \mapsto \left(\frac{a}{n}\right) is a Dirichlet character modulo nn. A value 11 does not make aa a square: (215)=1\left(\frac{2}{15}\right) = 1, but 22 is not a square modulo 33.

Theorem (quadratic reciprocity for the Jacobi symbol; Gauss, 1801; Jacobi, 1837). For odd, positive, coprime m,nm, n, (mn)(nm)=(−1)m−12n−12,(−1n)=(−1)n−12,(2n)=(−1)n2−18.\left(\frac{m}{n}\right)\left(\frac{n}{m}\right) = (-1)^{\frac{m-1}{2}\frac{n-1}{2}}, \qquad \left(\frac{-1}{n}\right) = (-1)^{\frac{n-1}{2}}, \qquad \left(\frac{2}{n}\right) = (-1)^{\frac{n^2-1}{8}}.

Definition (quadratic character). The quadratic character of a quadratic extension E/FE/F of number fields is the Hecke character η=ηE/F\eta = \eta_{E/F} of FF with η(t)=1\eta(t) = 1 if t∈F×NE/F(𝔸E×)t \in F^\times N_{E/F}(\mathbb{A}_E^\times) and η(t)=−1\eta(t) = -1 otherwise. By global class field theory (below), this subgroup has index 22 in 𝔸F×\mathbb{A}_F^\times, so η\eta has order 22. At an unramified place vv, ηv(ϖv)\eta_v(\varpi_v) is 11 if vv splits in EE and −1-1 if vv is inert.

Example. For F=ℚF = \mathbb{Q} and E=ℚ(−q)E = \mathbb{Q}(\sqrt{-q}) with q≡3(mod4)q \equiv 3 \pmod 4 prime, η(𝔭p)=(−qp)=(pq)\eta(\mathfrak p_p) = \left(\frac{-q}{p}\right) = \left(\frac{p}{q}\right) for each odd prime p≠qp \neq q.

IsChiField of VWin imposes (aq)=(ad)=1\left(\frac{a}{q}\right) = \left(\frac{a}{d}\right) = 1 on a∈(ℤ/Nℤ)×a \in (\mathbb{Z}/N\mathbb{Z})^\times. ResEtaPow of TLiftK asks that a Hecke character restricted to 𝔸F×\mathbb{A}_F^\times is ηn\eta^n, and quadraticClassFieldTheory of ThetaLocalInputs records the facts about η\eta that the proof uses.

Local class field theory

Let KK be a finite extension of ℚp\mathbb{Q}_p with residue field 𝔽q\mathbb{F}_q, maximal abelian extension KabK^{ab} and maximal unramified extension KurK^{ur}. Then Gal(Kur/K)≅ℤ^\mathrm{Gal}(K^{ur}/K) \cong \widehat{\mathbb{Z}}, generated by the Frobenius x↦xqx \mapsto x^q on residue fields. The Weil group WKW_K is the subgroup of Gal(K‾/K)\mathrm{Gal}(\overline K/K) of elements σ\sigma that act on KurK^{ur} as Frobdeg⁡σ\mathrm{Frob}^{\deg \sigma} with deg⁡σ∈ℤ\deg \sigma \in \mathbb{Z}. The inertia group IKI_K is the kernel of deg⁡\deg.

Theorem (local class field theory; Hasse, 1930; Lubin and Tate, 1965). There is a unique homomorphism ArtK:K×→Gal(Kab/K)\mathrm{Art}_K : K^\times \to \mathrm{Gal}(K^{ab}/K), the local Artin map, such that ArtK(ϖ)\mathrm{Art}_K(\varpi) acts on KurK^{ur} as the arithmetic Frobenius for each uniformizer ϖ\varpi, and NL/K(L×)N_{L/K}(L^\times) is the kernel of the induced map K×→Gal(L/K)K^\times \to \mathrm{Gal}(L/K) for each finite abelian L/KL/K. It is injective with dense image, maps 𝒪K×\mathcal{O}_K^\times onto the inertia subgroup, and induces WKab≅K×W_K^{ab} \cong K^\times. The map L↦NL/K(L×)L \mapsto N_{L/K}(L^\times) is a bijection from the finite abelian extensions of KK to the open subgroups of finite index of K×K^\times, with K×/NL/K(L×)≅Gal(L/K)K^\times / N_{L/K}(L^\times) \cong \mathrm{Gal}(L/K).

Many texts use the geometric normalization, which sends ϖ\varpi to a geometric Frobenius; it is the inverse.

Corollary. The characters of WKW_K are exactly the maps χ∘a\chi \circ a, with χ\chi a character of K×K^\times and a:WK→WKab≅K×a : W_K \to W_K^{ab} \cong K^\times the inverse of the Artin isomorphism.

IsLocalArtin in IsRecT is this map a:W→Ew×a : W \to E_w^\times, by five laws: onto (surj), continuous on inertia (cont), a(σ)a(\sigma) has valuation −deg⁡σ-\deg \sigma (ord_eq, so a geometric Frobenius goes to a uniformizer: the geometric normalization), each smooth character of WW is χ∘a\chi \circ a (factor), and compatible with each global Artin map of EE (glob). The step proves that exactly one such aa exists (artEx); the law one of IsRecT is the case GL1\mathrm{GL}_1.

Global class field theory

Let KK be a number field with maximal abelian extension KabK^{ab}.

Theorem (class field theory; Takagi, 1920; Artin, 1927; Chevalley, 1940). There is a unique continuous surjective homomorphism ArtK:CK→Gal(Kab/K)\mathrm{Art}_K : C_K \to \mathrm{Gal}(K^{ab}/K), the global Artin map, that restricts to the local Artin map on each Kv×K_v^\times. Its kernel is the identity component of CKC_K. The map L↦NL/K(CL)L \mapsto N_{L/K}(C_L) is a bijection from the finite abelian extensions LL of KK to the open subgroups of finite index of CKC_K, with CK/NL/K(CL)≅Gal(L/K)C_K / N_{L/K}(C_L) \cong \mathrm{Gal}(L/K).

At a real place that becomes complex in LL, Art\mathrm{Art} sends −1∈ℝ×-1 \in \mathbb{R}^\times to complex conjugation. At a finite place vv unramified in LL, Art(ϖv)\mathrm{Art}(\varpi_v) is the Frobenius element of vv in Gal(L/K)\mathrm{Gal}(L/K).

Definition (modulus). A modulus of KK is a formal product 𝔪=𝔪f𝔪∞\mathfrak m = \mathfrak m_f \mathfrak m_\infty of an ideal 𝔪f\mathfrak m_f of 𝒪K\mathcal{O}_K and a set 𝔪∞\mathfrak m_\infty of real places. Let IK𝔪I_K^{\mathfrak m} be the group of fractional ideals prime to 𝔪f\mathfrak m_f, and PK𝔪P_K^{\mathfrak m} its subgroup of principal ideals (α)(\alpha) with α≡1 mod ×𝔪f\alpha \equiv 1 \bmod^\times \mathfrak m_f and σ(α)>0\sigma(\alpha) > 0 for each σ∈𝔪∞\sigma \in \mathfrak m_\infty. The ray class group ClK𝔪=IK𝔪/PK𝔪\mathrm{Cl}_K^{\mathfrak m} = I_K^{\mathfrak m} / P_K^{\mathfrak m} is finite.

Theorem (Artin reciprocity; Artin, 1927). Let L/KL/K be finite abelian, and let the modulus 𝔪\mathfrak m be divisible by each finite prime and each real place that ramifies in LL. The Artin map IK𝔪→Gal(L/K)I_K^{\mathfrak m} \to \mathrm{Gal}(L/K), which sends each prime 𝔭∤𝔪\mathfrak p \nmid \mathfrak m to its Frobenius element, is surjective. If the exponents of 𝔪f\mathfrak m_f are large enough, its kernel is PK𝔪NL/K(IL𝔪)P_K^{\mathfrak m} N_{L/K}(I_L^{\mathfrak m}).

Definition (ray class field). For each modulus 𝔪\mathfrak m there is a unique abelian extension K𝔪/KK_{\mathfrak m}/K, the ray class field modulo 𝔪\mathfrak m, unramified outside 𝔪\mathfrak m, such that the Artin map induces ClK𝔪≅Gal(K𝔪/K)\mathrm{Cl}_K^{\mathfrak m} \cong \mathrm{Gal}(K_{\mathfrak m}/K). Each finite abelian extension lies in some ray class field.

Definition (Hilbert class field). The Hilbert class field HKH_K is the ray class field for 𝔪=1\mathfrak m = 1: the maximal abelian extension of KK unramified at each place. Then Gal(HK/K)≅ClK\mathrm{Gal}(H_K/K) \cong \mathrm{Cl}_K, and a prime 𝔭\mathfrak p splits completely in HKH_K if and only if it is principal.

Example. For K=ℚK = \mathbb{Q} and 𝔪=N∞\mathfrak m = N\infty, Clℚ𝔪≅(ℤ/Nℤ)×\mathrm{Cl}^{\mathfrak m}_{\mathbb{Q}} \cong (\mathbb{Z}/N\mathbb{Z})^\times and the ray class field is ℚ(ζN)\mathbb{Q}(\zeta_N). This gives the Kronecker-Weber theorem: each abelian extension of ℚ\mathbb{Q} lies in a cyclotomic field. For an imaginary quadratic KK, HK=K(j(E))H_K = K(j(E)) for an elliptic curve EE with complex multiplication by 𝒪K\mathcal{O}_K (Weber, 1908).

By Artin reciprocity, the characters of Gal(K𝔪/K)\mathrm{Gal}(K_{\mathfrak m}/K) are those of ClK𝔪\mathrm{Cl}^{\mathfrak m}_K, which are the finite order Hecke characters of conductor dividing 𝔪\mathfrak m.

UnitaryGroupThirdPlaceHypotheses states these theorems as globalArtinReciprocity, rayClassFieldSt, rayCharacterArtin and weberRayFieldSt, and the law glob of IsRecT uses the global Artin map. The field hj of the CM inputs says that a jj-invariant generates the Hilbert class field, and RayClassFieldSt is among the published arithmetic results.

Norms and Hermitian forms

Theorem (Hasse norm theorem; Hasse, 1931). Let L/KL/K be a cyclic extension of number fields and x∈K×x \in K^\times. Then xx is a norm from L×L^\times if and only if, at each place vv of KK, it is a norm from Lw×L_w^\times for one (so each) w∣vw \mid v.

This is a Hasse principle: a global property holds if it holds at each place. For quadratic E/FE/F, x∈F×x \in F^\times lies in N(E×)N(E^\times) if and only if ηv(x)=1\eta_v(x) = 1 for each vv.

Let E/FE/F be a quadratic extension with conjugation cc. A Hermitian form on an EE-vector space VV is a map h:V×V→Eh : V \times V \to E, EE-linear in the first variable, with h(y,x)=c(h(x,y))h(y, x) = c(h(x, y)). Its determinant det⁡h\det h is well defined in F×/N(E×)F^\times / N(E^\times). At a real place vv of FF that is complex in EE, hvh_v has a signature (pv,qv)(p_v, q_v).

Theorem (Landherr, 1936). Two nondegenerate Hermitian forms over E/FE/F of the same dimension are isomorphic if and only if they are isomorphic over Ev/FvE_v/F_v at each place vv of FF. A family of nondegenerate local forms hvh_v of one dimension nn comes from a global form if and only if ηv(det⁡hv)=1\eta_v(\det h_v) = 1 for almost all vv and ∏vηv(det⁡hv)=1\prod_v \eta_v(\det h_v) = 1.

Example. At a real place vv that is complex in EE, ηv(det⁡hv)=(−1)qv\eta_v(\det h_v) = (-1)^{q_v}. So the parities of the signatures of a global Hermitian form are tied to its finite places.

These results are LandherrSt and HasseNormSt in the published arithmetic results, and hermitianLocalGlobal and hermitianHassePrinciple in ThetaLocalInputs. With them the theta matching chapter builds the Hermitian spaces VV, V′V' and the skew-Hermitian space WW with given signatures.

Tate's thesis and Hecke L-functions

Let χ=∏vχv\chi = \prod_v \chi_v be a Hecke character of KK.

Definition (local factors). At a finite place, L(s,χv)=(1−χv(ϖv)qv−s)−1L(s, \chi_v) = (1 - \chi_v(\varpi_v) q_v^{-s})^{-1} if χv\chi_v is unramified, and 11 otherwise. At a real place with χv=sgnϵ∣⋅∣t\chi_v = \mathrm{sgn}^\epsilon |\cdot|^t, L(s,χv)=Γℝ(s+t+ϵ)L(s, \chi_v) = \Gamma_\mathbb{R}(s + t + \epsilon) with Γℝ(s)=π−s/2Γ(s/2)\Gamma_\mathbb{R}(s) = \pi^{-s/2}\Gamma(s/2). At a complex place with χv(z)=(z/∣z∣)m∣z∣2t\chi_v(z) = (z/|z|)^{m}|z|^{2t}, L(s,χv)=Γℂ(s+t+∣m∣/2)L(s, \chi_v) = \Gamma_\mathbb{C}(s + t + |m|/2) with Γℂ(s)=2(2π)−sΓ(s)\Gamma_\mathbb{C}(s) = 2(2\pi)^{-s}\Gamma(s). Chapter 3 defines Γ\Gamma.

Definition (Hecke L-function). The Hecke L-function of χ\chi is L(s,χ)=∏v∤∞L(s,χv)L(s, \chi) = \prod_{v \nmid \infty} L(s, \chi_v); for unitary χ\chi the product converges absolutely for Re⁡(s)>1\operatorname{Re}(s) > 1. For a finite set SS of places, the partial L-function LS(s,χ)L^S(s, \chi) omits the factors at v∈Sv \in S. The completed L-function is Λ(s,χ)=∏vL(s,χv)\Lambda(s, \chi) = \prod_v L(s, \chi_v) over all places.

Example. L(s,1)=ζK(s)=∑𝔞N(𝔞)−sL(s, 1) = \zeta_K(s) = \sum_{\mathfrak a} N(\mathfrak a)^{-s} is the Dedekind zeta function of KK; for K=ℚK = \mathbb{Q} it is the Riemann zeta function.

Definition (local epsilon factor). Let vv be finite, ψv\psi_v an unramified additive character, and χv\chi_v ramified with c=c(χv)c = c(\chi_v). The Gauss sum, up to the normalization of the text, is 𝔤(ψv,χv)=∑u∈𝒪v×/Uv(c)χv−1(u) ψv(u/ϖvc),\mathfrak g(\psi_v, \chi_v) = \sum_{u \in \mathcal{O}_v^\times / U_v^{(c)}} \chi_v^{-1}(u)\, \psi_v(u / \varpi_v^{c}), of absolute value qvc/2q_v^{c/2}. The epsilon factor ε(s,χv,ψv)\varepsilon(s, \chi_v, \psi_v) is 11 for unramified χv\chi_v, and for ramified χv\chi_v it is qv−csq_v^{-cs} times a constant built from 𝔤(ψv,χv)\mathfrak g(\psi_v, \chi_v), χv(ϖv)c\chi_v(\varpi_v)^{c} and powers of qvq_v. At each place, ε(s,χv,ψv)ε(1−s,χv−1,ψv)=χv(−1)\varepsilon(s, \chi_v, \psi_v)\varepsilon(1 - s, \chi_v^{-1}, \psi_v) = \chi_v(-1).

Theorem (Tate thesis; Hecke, 1918; Tate, 1950). Let χ\chi be a unitary Hecke character of KK and ψ\psi an additive character of 𝔸K\mathbb{A}_K. Then Λ(s,χ)\Lambda(s, \chi) extends to a meromorphic function on ℂ\mathbb{C}. It is entire unless χ=∣⋅∣𝔸it\chi = |\cdot|_{\mathbb{A}}^{i t} for a real tt, and then its only poles are simple, at s=−its = -it and s=1−its = 1 - it. It satisfies Λ(s,χ)=ε(s,χ) Λ(1−s,χ−1),ε(s,χ)=∏vε(s,χv,ψv)=W(χ) b1/2−s,\Lambda(s, \chi) = \varepsilon(s, \chi)\,\Lambda(1 - s, \chi^{-1}), \qquad \varepsilon(s, \chi) = \prod_v \varepsilon(s, \chi_v, \psi_v) = W(\chi)\,b^{1/2 - s}, with b=∣dK∣N(𝔣χ)b = |d_K| N(\mathfrak f_\chi) and ∣W(χ)∣=1|W(\chi)| = 1, independent of ψ\psi.

Tate integrates a Schwartz-Bruhat function against χ∣⋅∣s\chi |\cdot|^s over 𝔸K×\mathbb{A}_K^\times and applies Poisson summation for K⊂𝔸KK \subset \mathbb{A}_K; the local integrals give L(s,χv)L(s, \chi_v) and ε(s,χv,ψv)\varepsilon(s, \chi_v, \psi_v).

Corollary (Hecke). For χ≠∣⋅∣𝔸it\chi \neq |\cdot|_{\mathbb{A}}^{it}, L(s,χ)L(s, \chi) is entire and L(1,χ)≠0L(1, \chi) \neq 0. Each LS(s,χ)L^S(s, \chi) is L(s,χ)L(s, \chi) times finitely many polynomials in the qv−sq_v^{-s}, so it extends to ℂ\mathbb{C}.

The constants of ThetaIntegralsAgree, used in step 3 of the theta matching chapter, contain partial L-values LS0(s,ν)L^{S_0}(s, \nu) (charL) and ε(ψ,νv,z)=N(v)−c(νv)z 𝔤(ψ,νv) νv(ϖ)c(νv)\varepsilon(\psi, \nu_v, z) = N(v)^{-c(\nu_v) z}\, \mathfrak g(\psi, \nu_v)\, \nu_v(\varpi)^{c(\nu_v)}. The literature chapter samples the functional equation Λ(s)=ε∞(n)εf(χ,a,s)Λ′(1−s)\Lambda(s) = \varepsilon_\infty(n)\varepsilon_f(\chi, a, s)\Lambda'(1 - s) for a unitary Hecke character of a totally complex field with infinity type n≠0n \neq 0 relative to a CM type. heckeLPartialContinuation and heckeLEntireAndNonzeroAtOne of ThetaLocalInputs state the corollary.

Further reading

  • J. Neukirch, Algebraic Number Theory, chapters V (local class field theory), VI (global class field theory) and VII.8 (Hecke L-functions).
  • J. W. S. Cassels and A. Fröhlich (eds.), Algebraic Number Theory, chapters VI (Serre), VII (Tate) and XV (Tate's thesis).
  • W. Scharlau, Quadratic and Hermitian Forms, chapter 10.

The source of this page. Markdown. Something is wrong? Report it.