Finally, we state the theta lift from Definition 4.2.6 in the situation of the current lattice πΏ. Let π be a non-negative integer. Identifying H with H1 via the map π§ β¦βπ§π3, and using the identity from (4.3.5) and the fact that π(Im(π§π3)) = πIm(π§)2 for π§ β H we can write Shintaniβs theta function for the lattice πΏ as
ΞπΏ,π(π, π§) = πβπIm(π)1/2Im(π§)β2π βοΈ
π½βπΏβ²/πΏ
βοΈ
πβπΏ+π½
ππ(π§,1)ππ(π π(ππ§) +π π(ππ§β₯))eπ½ (4.3.6)
for π, π§ β H. Here the quantities π(ππ§) and π(ππ§β₯) are given as in Lemma 4.3.1. By Corollary 4.2.3 the theta functionΞπΏ,π(π, π§)is modular of weightπ+ 1/2with respect to ππΏ in π, and in the variable π§ it transforms as
ΞπΏ,π(π, πΎπ§) = π(πΎ, π§)2πΞπΏ,π(π, π§)
for π, π§ β H and πΎ β Ξ0(π). Accordingly, the (regularized) theta lift of a real analytic function πΉ:HβC[πΏβ²/πΏ] modular of weight π+ 1/2 with respect toππΏ is given by
Ξ¦πΏπ(π§;πΉ) = (πΉ,ΞπΏ,π(Β·, π§))reg
for π§ β H, and assuming that the (regularized) inner product exists the lift Ξ¦πΏπ(π§;πΉ) of πΉ is modular of weight 2π and level π.
For an integer π β₯ 2 the theta lift Ξ¦πΏπ(π§; Β·) is essentially the classical Shimura theta lift, mapping (vector valued) holomorphic cusp forms of weight π+ 1/2 to holomorphic cusp forms of weight 2π. It was introduced by Shimura in [Shi73] and later expressed as a theta lift by Niwa in [Niw75]. In particular, the Shimura lift of a holomorphic PoincarΓ© series is up to some constant the corresponding Zagier cusp form introduced in Section 2.4.2 (see for example [Oda77] or [KZ81]). More precisely, letπ½ βπΏβ²/πΏand πβZ+π(π½) with π >0. Then an easy unfolding argument shows that
Ξ¦πΏπ(π§;ππ+1/2,π½,π) = 2 (πβ1)!
ππ ππ,π½,4π π(π§) (4.3.7)
for π β Z with π β₯ 2 and π§ β H. Here ππ,π½,4π π(π§) is the meromorphic modular form introduced in Definition 2.4.6. Note that we can ignore the regularization of the inner product defining the theta lift in this case since ππ+1/2,π½,π(π) is a cusp form for π >0.
for π β π. Then (π, π) is a quadratic space of signature (2,2) with associated bilinear form given by (π, π) = βtr(ππ#) for π, π β π where (οΈπ π
π π
)οΈ#
= (οΈ π βπ
βπ π
)οΈ. Given a positive integer π we consider the lattice
πΏ:=
{οΈ(οΈ π π π π π
)οΈ
: π, π, π, πβZ }οΈ
inπ. ThenπΏ is an even lattice of level π with dual lattice given by πΏβ² =
{οΈ(οΈπ π/π
π π
)οΈ
: π, π, π, πβZ }οΈ
,
and the discriminant groupπΏβ²/πΏis isomorphic to the quotient(Z/πZ)2. It is well-known that the action of the identity component SO+(π) of the orthogonal group of π can be realized by the action of the group SL2(R)ΓSL2(R)on π(R)via
(π, πβ²).π :=πππβ²β1
for π, πβ² β SL2(R) and π β π(R). Here the four elements (Β±π,Β±πβ²) define the same element inSO+(π). Further, one easily checks that the discrete subgroupΞ0(π)ΓΞ0(π) of SL2(R)ΓSL2(R) acts on πΏβ², and that its subgroup Ξ(π)ΓΞ(π) fixes the classes of the discriminant group πΏβ²/πΏ, i.e., we have
Ξ(π)ΓΞ(π)βΞ(πΏ) and Ξ0(π)ΓΞ0(π)βSO+(πΏ), (4.4.1)
under the above identification of SO+(π)with the productSL2(R)ΓSL2(R). HereΞ(π) is the usual principal congruence subgroup given by
Ξ(π) :=
{οΈ(οΈπ π π π
)οΈ
βSL2(Z) :
(οΈπ π π π
)οΈ
β‘ (οΈ1 *
0 1 )οΈ
mod π }οΈ
, which clearly is a subgroup of Ξ0(π).
We now choose vectors π1 :=
(οΈ0 1 0 0
)οΈ
, π2 :=
(οΈ0 0 1 0
)οΈ
, π3 :=
(οΈ1 0 0 0
)οΈ
, π4 :=
(οΈ0 0 0 β1
)οΈ
.
Thenπ1 βπΏis primitive and isotropic andπ2 βπΏβ² is isotropic with(π1, π2) = 1. Moreover, the hyperbolic plane spanned by π1 and π2 is orthogonal to the plane spanned by the isotropic vectors π3 and π4 where (π3, π4) = 1. We define the sublattice πΎ =πΏβ©πβ₯1 β©πβ₯2, the corresponding quadratic spaceπ =πΎβQof signature(1,1)and the induced complex planeβ={π+ππ βπ(C) :π(π)>0}. ThenπΎ ={(οΈπ0
0 π
)οΈ:π, π βZ},π ={(οΈπ0
0π
)οΈ: π, πβ Q} and thus
β =
{οΈ(οΈπ§ 0 0 π§β²
)οΈ
: π§, π§β² βC with Im(π§) Im(π§β²)<0 }οΈ
.
Hence by choosing H2 as the connected component Hπ3βHπ4 of β we can identify the generalized upper half-planeH2 with the productHΓH. Letπ¦be the complex manifold defined in (4.1.1) associated to the space π(C). We chooseπ¦+ to be the image ofH2 in π¦ under the map π β¦β[ππΏ] where
ππΏ =π β(π(π) +π(π2))π1+π2 =
(οΈπ§ βπ§π§β² 1 βπ§β²
)οΈ
for π =π§π3+π§β²π4 βH2. Conversly, given π§, π§β² βH with π§=π₯+ππ¦, π§β² =π₯β² +ππ¦β² we set ππΏ(π§, π§β²) :=
(οΈπ₯ π¦π¦β²βπ₯π₯β² 1 βπ₯β²
)οΈ
and ππΏ(π§, π§β²) :=
(οΈπ¦ βπ₯π¦β²βπ₯β²π¦ 0 βπ¦β²
)οΈ
such thatππΏ(π§, π§β²) = Re(ππΏ),ππΏ(π§, π§β²) = Im(ππΏ)forπ =π§π3+π§β²π4. Thenπ(ππΏ(π§, π§β²)) = π(ππΏ(π§, π§β²)) =π¦π¦β², and the identification ofH2 with the Grassmannian Gr(π) is given by the map π§π3+π§β²π4 β¦βRππΏ(π§, π§β²)βRππΏ(π§, π§β²).
In order to understand the action of the groupSL2(R)ΓSL2(R) onH2 we compute π((π, πβ²), π) = ((π, πβ²).ππΏ, π1) =π(π, π§)π(πβ², π§β²)
(4.4.2)
for π, πβ² β SL2(R) and π = π§π3 +π§β²π4 β H2. Here the automorphy factors on the right are the classical ones defined at the beginning of Section 2.4. Hence we find
((π, πβ²).π)πΏ=π((π, πβ²), π)β1(π, πβ²).ππΏ=
(οΈππ§ βππ§Β·πβ²π§β² 1 βπβ²π§β²
)οΈ
= (ππ§ π3 +πβ²π§β²π4)πΏ (4.4.3)
for π, πβ² β SL2(R) and π = π§π3 +π§β²π4 β H2, i.e., (π, πβ²).π = ππ§ π3+πβ²π§β²π4. Therefore, the identification of H2 with the product HΓH is SL2(R)ΓSL2(R)-equivariant where the latter group naturally acts on HΓH via(π, πβ²).(π§, π§β²) = (ππ§, πβ²π§β²).
In particular, for π β Z a function πΉ: H2 β C is modular of weight π with respect to the group Ξ(π)ΓΞ(π) in the sense of Definition 4.1.2 if and only if the function π(π§, π§β²) = πΉ(π§π3 +π§β²π4) for π§, π§β² β H is modular of weight π for the group Ξ(π) in the classical sense in both variables, i.e., if
π(πΎπ§, πΎβ²π§β²) = π(πΎ, π§)ππ(πΎβ², π§β²)ππ(π§, π§β²) for all πΎ, πΎβ² βΞ(π) and π§, π§β² βH.
Forπβπ and π§, π§β² βH we denote the orthogonal projection of π onto the positive or negative definite subspace RππΏ(π§, π§β²)βRππΏ(π§, π§β²) orππΏ(π§, π§β²)β₯β©ππΏ(π§, π§β²)β₯ in π(R) by ππ§,π§β² orπ(π§,π§β²)β₯, respectively. As π(Im(π)) = Im(π§) Im(π§β²)we can restate equation (4.1.2) in the present setting as
π(ππ§,π§β²) = |(π, ππΏ)|2 4 Im(π§) Im(π§β²) (4.4.4)
for π β π and π§, π§β² β H with π = π§π3 +π§β²π4. Further, given π§, π§β² β H we denote the majorant of π associated to the pair (π§, π§β²) as defined in (4.1.3) by
ππ§,π§β²(π) =π(ππ§,π§β²)βπ(π(π§,π§β²)β₯), and using (4.4.4) we thus find
ππ§,π§β²(π) = |(π, ππΏ)|2
2 Im(π§) Im(π§β²)βπ(π) (4.4.5)
for πβπ and π§, π§β² βH with π =π§π3+π§β²π4. 86
Eventually, we present the theta liftΞ¦πΏπ(π;πΉ)from Section 4.2 in the case ofπΏbeing our current lattice of signature(2,2). Recall thatπ(Im(π)) = Im(π§) Im(π§β²)for π =π§π3+π§β²π4 with π§, π§β² β H. Let now π be a non-negative integer. Identifying H2 with the product HΓH we can write Shintaniβs theta function for the latticeπΏ as
ΞπΏ,π(π, π§, π§β²) = π£ Im(π§)βπIm(π§β²)βπ βοΈ
π½βπΏβ²/πΏ
βοΈ
πβπΏ+π½
(π, ππΏ)ππ(οΈ
π’π(π) +ππ£ππ§,π§β²(π))οΈ
eπ½ (4.4.6)
forπ, π§, π§β² βHwith π =π’+ππ£. The theta functionΞπΏ,π(π, π§)is modular of weightπ with respect to ππΏ inπ, and in the variables π§ and π§β² it transforms as
ΞπΏ,π(π, πΎπ§, πΎβ²π§β²) = π(πΎ, π§)ππ(πΎβ², π§β²)πΞπΏ,π(π, π§, π§β²)
for π, π§, π§β² βH and πΎ, πΎβ² βΞ(π). Therefore, the (regularized) theta lift of a real analytic function πΉ:HβC[πΏβ²/πΏ] modular of weight π with respect toππΏ is given by
Ξ¦πΏπ(π§, π§β²;πΉ) = (πΉ,ΞπΏ,π(Β·, π§, π§β²))reg
for π§, π§β² βH, and the lift Ξ¦πΏπ(π§, π§β²;πΉ) of πΉ is modular of weight π for the group Ξ(π)in both variables π§ and π§β² whenever the (regularized) inner product exists. In this setting, the lift Ξ¦πΏπ(π§, π§β²;πΉ) is sometimes called the Doi-Naganuma lift of πΉ, which classically is a lift from scalar valued modular forms to Hilbert modular forms (see [DN67, DN69]).
5 Realizing non-holomorphic
Eisenstein series as theta lifts
In this chapter we realize the non-holomorphic Eisenstein series introduced in Section 2.6 as regularized theta lifts in two different ways. On the one hand we obtain averaged versions of these Eisenstein series as the theta lift of Selbergβs PoincarΓ© series of the first kind using the lattice of signature (2,1)from Section 4.3. Remarkably, we indeed obtain averaged versions of hyperbolic, parabolic an elliptic Eisenstein series as the theta lift of a single type of PoincarΓ© series.
On the other hand we realize the hyperbolic kernel function defined in Section 2.6.4 as the theta lift of a more elementary PoincarΓ© series introduced in Section 3.6, using the lattice of signature (2,2) from Section 4.4. We can then use this representation of the hyperbolic kernel function to obtain individual hyperbolic, parabolic and elliptic Eisenstein series as variations of this theta lift.
5.1 Regularized theta lifts of non-holomorphic PoincarΓ© series
Let (π, π) be a quadratic space of signature (2, π) and let πΏ be an even lattice in π. We further fix vectors π1 β πΏ and π2 β πΏβ² such that π1 is primitive with π(π1) = 0 and (π1, π2) = 1 as in Section 4.1, and we let Hπ be the induced generalized upper half-plane.
Forπ½ βπΏβ²/πΏ and πβZ+π(π½) we set
πΏπ½,π:={π βπΏ+π½: π(π) = π}, and we define the subset π»π½,ππΏ of Hπ by
π»π½,ππΏ := βοΈ
πβπΏπ½,π πΜΈ=0
{π βHπ: (π, ππΏ) = 0}.
(5.1.1)
Note thatπ»π½,ππΏ =β ifπβ₯0sinceπβπΏπ½,πcan only be orthogonal to someππΏ=ππΏ+πππΏ with π βHπ ifπ= 0 orπ(π)<0as RππΏβRππΏ defines a2-dimensional positive definite subspace of π(R).
Letπ be a non-negative integer and setπ := 1 +πβπ/2. Then π satisfies the condition given in (3.4.1). In the present section we study the regularized theta lift of two of the non-holomorphic PoincarΓ© series given in Definition 3.4.3, namely Selbergβs PoincarΓ© series of the first kind ππ ,π½,ππΏ (π, π )and the simpler, non-standard PoincarΓ© series ππΏπ ,π½,π(π, π ).
Definition 5.1.1. Letπ βZ with πβ₯0, π = 1 +πβπ/2,π½ βπΏβ²/πΏ and πβZ+π(π½).
(a) We define the regularized theta lift of Selbergβs PoincarΓ© series of the first kind ππ ,π½,ππΏ (π, π ) by
Ξ¦Sel,πΏπ,π½,π(π, π ) = Ξ¦πΏπ(π;ππ ,π½,ππΏ (Β·, π )) for π βHπβπ»π½,ππΏ and π βCwith Re(π )>1 +π/2.
(b) We define the regularized theta lift of the PoincarΓ© series ππΏπ ,π½,π(π, π )by Ξ¦Q,πΏπ,π½,π(π, π ) = Ξ¦πΏπ(π;ππΏπ ,π½,π(Β·, π ))
for π βHπ and π βC with Re(π )>1 +π/2.
Here Ξ¦πΏπ(π; Β·) denotes the regularized theta lift from Definition 4.2.6. We will see in Theorem 5.1.3 that the above definition is indeed well-defined, i.e., that the corresponding (regularized) integrals exist in all the given cases. Thus,Ξ¦Sel,πΏπ,π½,π(π, π )and Ξ¦Q,πΏπ,π½,π(π, π ) are modular of weight π in the variableπ with respect to the groupΞ(πΏ) defined in (4.1.4).
Before we start studying the above lifts, we quickly recall some technicalities. For π βHπ the majorant of the present quadratic formπ associated toπ is given byππ(π) = π(ππ)βπ(ππβ₯)forπ βπ (see equation (4.1.3)). Hereππ and ππβ₯ denote the orthogonal projections of πonto the positive or negative definite subspaces ofπ corresponding to π and πβ₯, respectively. The majorant ππ defines a positive definite quadratic form on π, and thus the Epstein zeta function associated to ππ and the lattice πΏβ² defined by
π(π ;ππ, πΏβ²) := βοΈ
πβπΏβ²β{0}
ππ(π)βπ (5.1.2)
converges absolutely and locally uniformly for π β C with Re(π ) > 1 +π/2 (see for example [Eps03]).
Lemma 5.1.2. Let π½ βπΏβ²/πΏ, πβZ+π(π½) and π βHπβπ»π½,ππΏ . Then the sum
βοΈ
πβπΏπ½,πβ{0}
π(ππ)βπ
converges absolutely and locally uniformly for π βC with Re(π )>1 +π/2.
Proof. Firstly, we note that for π β πΏπ½,π we have π(ππ) = 0 if and only if (π, ππΏ) = 0 by (4.1.2), which is the case if and only if π = 0 or π β π»π½,ππΏ . Thus the given sum is well-defined. Next, we can choose a finite subset πΎ β πΏβ² such that ππ(π)β₯ 2|π| for all πβπΏβ²βπΎ since ππ is positive definite. Hence we obtain
π(ππ) = ππ(π) +π
2 = ππ(π)
4 + ππ(π) + 2π
4 β₯ ππ(π) 4 for πβπΏπ½,πβπΎ, giving
βοΈ
πβπΏπ½,π πΜΈ=0
ββπ(ππ)βπ β
ββ€ βοΈ
πβπΏπ½,πβ©πΎ πΜΈ=0
π(ππ)βRe(π )+ 4Re(π ) βοΈ
πβπΏπ½,πβπΎ πΜΈ=0
ππ(π)βRe(π ).
Here the first sum is finite and the second sum is dominated by the Epstein zeta function associated toππ and πΏβ² given in (5.1.2), which converges absolutely and locally uniformly for π βC with Re(π )>1 +π/2.
90
We can now prove that the theta lifts given in Definition 5.1.1 are indeed well-defined.
Moreover, if π >0 we already obtain a meromorphic continuation of the two lifts to the half-plane defined by Re(π )>1 +π/2βπ/2
Theorem 5.1.3. Let π βZ with π β₯0, π = 1 +πβπ/2, π½ βπΏβ²/πΏ and πβZ+π(π½).
(a) The regularized theta lift of Selbergβs PoincarΓ© seriesππ ,π½,ππΏ (π, π )defines a real analytic function in π β Hπβπ»π½,ππΏ and a holomorphic function in π for Re(π ) > 1 +π/2, which is given by
Ξ¦Sel,πΏπ,π½,π(π, π ) = 2 Ξ(π +π) 4π ππ +π
βοΈ
πβπΏπ½,π πΜΈ=0
π(ππ)βπ (π, ππΏ)βπ.
For π >0 and π β Hπβπ»π½,ππΏ the sum on the right-hand side yields a holomorphic continuation of Ξ¦Sel,πΏπ,π½,π(π, π ) in π to the half-plane Re(π )>1 +π/2βπ/2.
(b) The regularized theta lift of the PoincarΓ© series ππΏπ ,π½,π(π, π ) defines a real analytic function in π β Hπ and a holomorphic function in π for Re(π )> 1 +π/2, which is given by
Ξ¦Q,πΏπ,π½,π(π, π ) = 2 Ξ(π +π)
(2π)π +π π(Im(π))βπ βοΈ
πβπΏπ½,π
πΜΈ=0
ππ(π)βπ βπ(π, ππΏ)π.
For π >0 and π βHπ the sum on the right-hand side yields a holomorphic continu-ation of Ξ¦Q,πΏπ,π½,π(π, π ) in π to the half-plane Re(π )>1 +π/2βπ/2.
Proof. Since (a) and (b) are poven analogously, we only give a detailed proof for part (a), and afterwards comment on the necessary adaptions of the given proof in the case of (b).
For (a) fix π β Hπ βπ»π½,ππΏ and π β C with Re(π ) > 1 + π/2. Evaluating the inner product β¨οΈ
ππ ,π½,ππΏ (π, π ),ΞπΏ,π(π, π)β©οΈ
and splitting the sum over matrices π β β¨πβ©βSL2(Z) coming from the PoincarΓ© series into matrices π with lower left entry πΜΈ= 0 and π= 0 we can write the theta lift Ξ¦Sel,πΏπ,π½,π(π, π ) as
CTπ‘=0 [οΈ
πlimββ
β«οΈ
β±π
βοΈ
πββ¨πβ©βSL2(Z) πΜΈ=Β±1
Im(π π)π +π π(ππ π)ΞπΏ,π,π½(π π, π)π£βπ‘ππ(π) ]οΈ
(5.1.3)
+ 2 CTπ‘=0 [οΈ
πlimββ
β«οΈ
β±π
π£π +π βπ‘π(ππ)ΞπΏ,π,π½(π, π)ππ(π) ]οΈ
. Here we use the notationπ =π’+ππ£,π = 1 +πβπ/2and ΞπΏ,π,π½(π, π) = β¨ΞπΏ,π(π, π),eπ½β©.
Letβ:={π βH: |Re(π)| β€1/2}. Then β is a rectangular fundamental domain for the action of β¨πβ© on H, and by Corollary 4.2.5 we have
β«οΈ
βββ±
β
β
βπ£π +π π(ππ)ΞπΏ,π,π½(π, π)
β
β
βππ(π)β€πΆ
β«οΈ 1 0
π£Re(π )βπ/2β2
ππ£ <β (5.1.4)
for some constantπΆ > 0asRe(π )>1 +π/2. Hence we can plug inπ‘= 0 in the first term in (5.1.3), take the limit π β β, interchange summation and integration and apply the usual unfolding trick to find
CTπ‘=0 [οΈ
πlimββ
β«οΈ
β±π
βοΈ
πββ¨πβ©βSL2(Z) πΜΈ=Β±1
Im(π π)π +π π(ππ π)ΞπΏ,π,π½(π π, π)π£βπ‘ππ(π) ]οΈ
= 2
β«οΈ
βββ±
π£π +π π(ππ)ΞπΏ,π,π½(π, π)ππ(π).
In the second term in (5.1.3) we split the integral at the horizontal line π£ = 1, i.e., CTπ‘=0
[οΈ
πlimββ
β«οΈ
β±π
π£π +π βπ‘π(ππ)ΞπΏ,π,π½(π, π)ππ(π) ]οΈ
=
β«οΈ
β±1
π£π +π π(ππ)ΞπΏ,π,π½(π, π)ππ(π) + CTπ‘=0
[οΈ
πlimββ
β«οΈ
β±πββ±1
π£π +π βπ‘π(ππ)ΞπΏ,π,π½(π, π)ππ(π) ]οΈ
. Here we can drop the regularization in the first term on the right-hand side as β±1 is compact. Putting everything back together we obtain
Ξ¦Sel,πΏπ,π½,π(π, π ) = 2
β«οΈ 1 0
β«οΈ 1/2
β1/2
π£π +π π(ππ)ΞπΏ,π,π½(π, π)ππ’ππ£ π£2 (5.1.5)
+ 2 CTπ‘=0 [οΈβ«οΈ β
1
β«οΈ 1/2
β1/2
π£π +π βπ‘π(ππ)ΞπΏ,π,π½(π, π)ππ’ππ£ π£2
]οΈ
. In both terms the integral with respect to π’ is simply given by
β«οΈ 1/2
β1/2
π(ππ’)ΞπΏ,π,π½(π, π)ππ’.
Plugging in the definition of the π½-th component of the theta function ΞπΏ,π(π, π) and using that
βοΈ
πβπΏ+π½
β«οΈ 1/2
β1/2
β
β
βπ(ππ’)(π, ππΏ)ππ(π π(ππ) +π π(ππβ₯))
β
β
βππ’= βοΈ
πβπΏ+π½
ββ(π, ππΏ)ππβ2ππ£ππ(π)β
β<β as the sum defining ΞπΏ,π,π½(π, π)is absolutely convergent, we obtain
β«οΈ 1/2
β1/2
π(ππ’)ΞπΏ,π,π½(π, π)ππ’ = π£π/2 π(Im(π))π
βοΈ
πβπΏ+π½
(π, ππΏ)ππβ2ππ£ππ(π)
β«οΈ 1/2
β1/2
π(π’(πβπ(π))ππ’.
Hereππ(π) = π(ππ)βπ(ππβ₯) as in (4.1.3), and the integral on the right-hand side is1 if π(π) = π and vanishes otherwise. Therefore, we can write (5.1.5) as
Ξ¦Sel,πΏπ,π½,π(π, π ) = 2 π(Im(π))π
(οΈβ«οΈ 1 0
π£π +πβ1 βοΈ
πβπΏπ½,π
(π, ππΏ)ππβ4ππ£π(ππ)ππ£
+ CTπ‘=0 [οΈβ«οΈ β
1
π£π +πβπ‘β1 βοΈ
πβπΏπ½,π
(π, ππΏ)ππβ4ππ£π(ππ)ππ£ ]οΈ)οΈ
.
92
Here the contribution of the element π= 0 appearing only if π =π½ = 0 is zero since
β«οΈ 1 0
π£π +πβ1ππ£+ CTπ‘=0
β«οΈ β 1
π£π +πβπ‘β1ππ£= 1
π +π βCTπ‘=0
(οΈ 1 π +πβπ‘
)οΈ
= 0.
Hence we can exclude the element π = 0 in each of the two sums above. Moreover, we reunite the remaining terms, giving
Ξ¦Sel,πΏπ,π½,π(π, π ) = 2
π(Im(π))πCTπ‘=0 [οΈβ«οΈ β
0
π£π +πβπ‘β1 βοΈ
πβπΏπ½,π πΜΈ=0
(π, ππΏ)ππβ4ππ£π(ππ)ππ£ ]οΈ
. (5.1.6)
Next we note that π(ππ)>0 and (π, ππΏ)ΜΈ= 0 for πβπΏπ½,π with πΜΈ= 0 asπ βHπβπ»π½,ππΏ . Thus, using equation (4.1.2) we find that
βοΈ
πβπΏπ½,π πΜΈ=0
β«οΈ β 0
β
β
βπ£π +πβ1(π, ππΏ)ππβ4ππ£π(ππ)
β
β
βππ£= Ξ(Re(π ) +π)π(Im(π))π/2 4Re(π )+π/2πRe(π )+π
βοΈ
πβπΏπ½,π πΜΈ=0
π(ππ)βRe(π )βπ/2 (5.1.7)
forRe(π )+π >0, and by Lemma 5.1.2 the sum on the right-hand side converges absolutely and locally uniformly for Re(π )>1 +π/2βπ/2. Therefore we can simply plug in π‘= 0 on the right-hand side of (5.1.6), interchange summation and integration and use again (4.1.2) to proof the statement given in part (a) of the theorem.
In order to also proof part (b) we need to modify the above proof of (a) as follows: As in equation (5.1.6) we obtain
Ξ¦Q,πΏπ,π½,π(π, π ) = 2
π(Im(π))πCTπ‘=0 [οΈβ«οΈ β
0
π£π +πβπ‘β1 βοΈ
πβπΏπ½,π
πΜΈ=0
(π, ππΏ)ππβ2ππ£ππ(π)ππ£ ]οΈ
(5.1.8)
for π β Hπ (without the exclusion of π»π½,ππΏ ) and π β C with Re(π ) >1 +π/2, where in (5.1.6) we find the term πβ4ππ£π(ππ) = πβ2ππ£ππ(π)πβ2ππ£π instead of πβ2ππ£ππ(π) as in (5.1.8).
Here the missing factor, namely πβ2ππ£π, is exactly the term by which the two building blocks π£π π(ππ)eπ½ and π£π π(ππ’)eπ½ defining the PoincarΓ© seriesππ ,π½,ππΏ (π, π )andππΏπ ,π½,π(π, π ) differ. Now as in equation (5.1.7) we see that
βοΈ
πβπΏπ½,π πΜΈ=0
β«οΈ β 0
β
β
βπ£π +πβ1(π, ππΏ)ππβ2ππ£ππ(π)
β
β
βππ£ (5.1.9)
= Ξ(Re(π ) +π)π(Im(π))π/2 2Re(π )πRe(π )+π
βοΈ
πβπΏπ½,π πΜΈ=0
π(ππ)π/2ππ(π)βRe(π )βπ,
where we have used again the identity (4.1.2). As π(ππ)β€ππ(π) = π(ππ) +|π(ππβ₯)|, and since the sumβοΈ
πβπΏβ²β{0}ππ(π)βπ converges forRe(π )>1 +π/2as remarked in (5.1.2), we find that the right-hand side of (5.1.9) converges for Re(π )> 1 +π/2βπ/2. Therefore we can plug in π‘ = 0 on the right-hand side of (5.1.8), interchange summation and integration, and compute the remaining integral. This finishes the proof of part (b).
Remark 5.1.4.
(1) Ifπ½ =βπ½inπΏβ²/πΏand2βπβ2π β‘2mod4then the PoincarΓ© seriesππ ,π½,ππΏ andππΏπ ,π½,π vanish identically. This matches the two formulas given in the previous theorem since if π½ = βπ½ then πΏπ½,π = βπΏπ½,π and thus the sums in (a) and (b) of Theorem 5.1.3 cancel completely if π is odd as
π((βπ)π) =π(ππ), ππ(βπ) =ππ(π), (βπ, ππΏ) = β(π, ππΏ)
for πβπ and π βHπ. On the other hand, the congruence 2βπβ2π β‘2 mod4 is satisfied if and only if the corresponding non-negative integerπ(withπ = 1+πβπ/2) is odd.
(2) Ifπ = 0 thenππ ,π½,0πΏ (π, π ) = ππΏπ ,π½,0(π, π ), and thus also the corresponding lifts need to agree, i.e., we have
Ξ¦Sel,πΏπ,π½,0(π, π ) = Ξ¦Q,πΏπ,π½,0(π, π )
for π β Hπ and π β C with Re(π ) >1 +π/2βπ/2. This agrees with the formulas given in Theorem 5.1.3 since
π(ππ) = ππ(π)
2 and (π, ππΏ) = 2ππ(π)π(Im(π)) (π, ππΏ)
for πβπΏπ½,0 with πΜΈ= 0 and π βHπ. Here the first equality is clear asπ(π) = 0, and the second equality follows from the identity in (4.1.2).
In the following two sections we now evaluate the above theta lifts for the special lattices of signature (2,1) and (2,2) from Section 4.3 and Section 4.4, respectively. Though we are mainly interested in the lift of Selbergβs PoincarΓ© series in the case of signature(2,1), and in the lift of the PoincarΓ© series ππΏπ ,π½,π(π, π ) in the case of signature (2,2), we also present the opposite cases for the sake of completeness.