Waldspurger formula

In representation theory, a branch of mathematics, the Waldspurger formula relates the special values of two L-functions of two related admissible irreducible representations. Let k be the base field, f be an automorphic form over k, π be the representation associated via the Jacquet–Langlands correspondence with f. Goro Shimura (1976) proved this formula when k=ℚ and f is a cusp form; Günter Harder made the same discovery at the same time in an unpublished paper. Marie-France Vignéras (1980) proved this formula when k=ℚ and f is a newform. Jean-Loup Waldspurger, for whom the formula is named, reproved and generalized the result of Vignéras in 1985 via a totally different method which was widely used thereafter by mathematicians to prove similar formulas.

Statement

Let k be a number field, 𝔸 be its adele ring, k× be the subgroup of invertible elements of k, 𝔸× be the subgroup of the invertible elements of 𝔸, χ,χ1,χ2 be three quadratic characters over 𝔸×/k×, G=SL2(k), 𝒜(G) be the space of all cusp forms over G(k)∖G(𝔸), ℋ be the Hecke algebra of G(𝔸). Assume that, π is an admissible irreducible representation from G(𝔸) to 𝒜(G), the central character of π is trivial, πν∼π[hν] when ν is an archimedean place, A is a subspace of 𝒜(G) such that π|ℋ:ℋ→A. We suppose further that, ε(π⊗χ,1/2) is the Langlands ε-constant [ (Langlands 1970); (Deligne 1972) ] associated to π and χ at s=1/2. There is a γ∈k× such that k(χ)=k(γ).

Definition 1. The Legendre symbol (χπ)=ε(π⊗χ,1/2)⋅ε(π,1/2)⋅χ(−1).

Definition 2. Let Dχ be the discriminant of χ. p(χ)=Dχ1/2∑ν archimedean|γν|νhν/2.

Definition 3. Let f0,f1∈A. b(f0,f1)=∫x∈k×f0(x)⋅f1(x)‾dx.

Definition 4. Let T be a maximal torus of G, Z be the center of G, φ∈A. β(φ,T)=∫t∈Z∖Tb(π(t)φ,φ)dt.

Let K be a field such that k(π)⊂K⊂ℂ. One can choose a K-subspace A0 of A such that (i) A=A0⊗Kℂ; (ii) (A0)π(G)=A0. De facto, there is only one such A0 modulo homothety. Let T1,T2 be two maximal tori of G such that χT1=χ1 and χT2=χ2. We can choose two elements φ1,φ2 of A0 such that β(φ1,T1)≠0 and β(φ2,T2)≠0.

Definition 5. Let D1,D2 be the discriminants of χ1,χ2.

p(π,χ1,χ2)=D1−1/2D21/2L(χ1,1)−1L(χ2,1)L(π⊗χ1,1/2)L(π⊗χ2,1/2)−1β(φ1,T1)−1β(φ2,T2).

We take Σf to be the set {all finite k-places ν∣πν doesn't map nonzero vectors invariant under the action of GL2(kν) to zero}, Σs to be the set of (all k-places ν∣ν is real, or finite and special).

Theorem —Let k=ℚ. We assume that, (i) L(π⊗χ2,1/2)≠0; (ii) for ν∈Σs, (χ1,νπν)=(χ2,νπν) . Then there is a constant q∈ℚ(π) such that L(π⊗χ1,1/2)L(π⊗χ2,1/2)−1=qp(χ1)p(χ2)−1∏ν∈Σfp(πν,χ1,ν,χ2,ν)

Comments:

  1. The formula in the theorem is the well-known Waldspurger formula. It is of global-local nature, in the left with a global part, in the right with a local part. By 2017, mathematicians often call it the classic Waldspurger's formula.
  2. It is worthwhile to notice that when the two characters are equal, the formula can be greatly simplified.
  3. [ (Waldspurger 1985), Thm 6, p. 241 ] When one of the two characters is 1, Waldspurger's formula becomes much simpler. Without loss of generality, we can assume that, χ1=χ and χ2=1. Then there is an element q∈ℚ(π) such that L(π⊗χ,1/2)L(π,1/2)−1=qDχ1/2.

The case when Fp(T) and φ is a metaplectic cusp form

Let p be prime, 𝔽p be the field with p elements, R=𝔽p[T],k=𝔽p(T),k∞=𝔽p((T−1)),o∞ be the integer ring of k∞,ℋ=PGL2(k∞)/PGL2(o∞),Γ=PGL2(R). Assume that, N,D∈R, D is squarefree of even degree and coprime to N, the prime factorization of N is ∏ℓℓαℓ. We take Γ0(N) to the set {(abcd)∈Γ∣c≡0modN}, S0(Γ0(N)) to be the set of all cusp forms of level N and depth 0. Suppose that, φ,φ1,φ2∈S0(Γ0(N)).

Definition 1. Let (cd) be the Legendre symbol of c modulo d, SL~2(k∞)=Mp2(k∞). Metaplectic morphism η:SL2(R)→SL~2(k∞),(abcd)↦((abcd),(cd)).

Definition 2. Let z=x+iy∈ℋ,dμ=dxdy|y|2. Petersson inner product ⟨φ1,φ2⟩=[Γ:Γ0(N)]−1∫Γ0(N)∖ℋφ1(z)φ2(z)‾dμ.

Definition 3. Let n,P∈R. Gauss sum Gn(P)=∑r∈R/PR(rP)e(rnT2).

Let λ∞,φ be the Laplace eigenvalue of φ. There is a constant θ∈ℝ such that λ∞,φ=e−iθ+eiθp.

Definition 4. Assume that v∞(a/b)=deg⁡(a)−deg⁡(b),ν=v∞(y). Whittaker function W0,iθ(y)={peiθ−e−iθ[(eiθp)ν−1−(e−iθp)ν−1],when ν≥2;0,otherwise.

Definition 5. Fourier–Whittaker expansion φ(z)=∑r∈Rωφ(r)e(rxT2)W0,iθ(y). One calls ωφ(r) the Fourier–Whittaker coefficients of φ.

Definition 6. Atkin–Lehner operator Wαℓ=(ℓαℓbNℓαℓd) with ℓ2αℓd−bN=ℓαℓ.

Definition 7. Assume that, φ is a Hecke eigenform. Atkin–Lehner eigenvalue wαℓ,φ=φ(Wαℓz)φ(z) with wαℓ,φ=±1.

Definition 8. L(φ,s)=∑r∈R∖{0}ωφ(r)|r|ps.

Let S~0(Γ~0(N)) be the metaplectic version of S0(Γ0(N)), {E1,…,Ed} be a nice Hecke eigenbasis for S~0(Γ~0(N)) with respect to the Petersson inner product. We note the Shimura correspondence by Sh.

Theorem [ (Altug & Tsimerman 2010), Thm 5.1, p. 60 ]. Suppose that Kφ=1p(p−e−iθ)(p−eiθ), χD is a quadratic character with Δ(χD)=D. Then ∑Sh⁡(Ei)=φ|ωEi(D)|p2=KφG1(D)|D|p−3/2⟨φ,φ⟩L(φ⊗χD,1/2)∏ℓ(1+(ℓαℓD)wαℓ,φ).