Biextensions of Modular Motives

• Biextensions of modular motives are mixed motivic objects that encode two compatible extension structures from elliptic curves and cusp forms.
• They are constructed via explicit modular cohomology sequences involving symmetric powers of H¹ and relative cohomology on modular curves.
• This framework offers a motivic interpretation of Gross–Zagier log-algebraicity, linking single-valued periods with regulators and special L-values.

Searching arXiv for papers on biextensions of modular motives and related motivic period constructions. Biextensions of modular motives are mixed motivic objects that simultaneously encode two compatible extension structures attached to modular cohomology, typically involving symmetric powers of the motive of an elliptic curve and cusp-form motives. In the setting of modular curves and the universal elliptic curve, they arise from relative cohomology with marked points and admit realizations in a Tannakian category of compatible Betti and de Rham data with real Frobenius. Their single-valued periods are matrix-valued higher Green’s functions on the upper half-plane, and in the complex multiplication (CM) setting, after Hecke projection, they produce Kummer-type extensions whose single-valued period is the logarithm of an algebraic number. This furnishes a motivic interpretation of the Gross–Zagier log-algebraicity conjecture and, in level $1$ and weight $4$, a geometric proof via the mixed Tate property of $\mathcal{M}_{1,3}$ (Brown et al., 6 Aug 2025).

1. Conceptual framework

In an abelian tensor category, a biextension of $(M,N)$ by $P$ is an object $E$ equipped with compatible exact sequences

$$0 \to P \to E \to M \to 0, \qquad 0 \to M \to E \to N \to 0,$$

together with a bilinear structure. Under the usual bilinearity condition, biextension classes are related to extension groups by

$$\operatorname{Biext}(M,N;P)\cong \operatorname{Ext}^1(M\otimes N,P).$$

In the theory of mixed Hodge structures and 1-motives, this formalism goes back to Deligne’s theory, including the Poincaré biextension.

For modular motives, the relevant realization category is a neutral Tannakian category $H_K$ over $\mathbf{Q}$, whose objects are tuples

$4$0

Within this category, biextensions are not merely abstract extension classes: they are realized by explicit modular-cohomological exact sequences coming from punctured modular curves with marked points. The key feature is that the resulting mixed object carries arithmetic information simultaneously in two variables, corresponding to two marked points on the modular curve.

A structural input is the polarization on $4$1 for an elliptic curve $4$2. The canonical antisymmetric pairing

$4$3

induces $4$4, hence

$4$5

This is central in the CM case, where symmetric powers can contain Tate summands, enabling passage from modular biextensions to Kummer-type extensions (Brown et al., 6 Aug 2025).

2. Modular motives from elliptic curves and marked points

Let $4$6 be a subgroup of finite index, and let $4$7 be the corresponding modular curve, viewed as a smooth stack with analytification $4$8. Let

$4$9

be the universal elliptic curve. The basic coefficient system is

$\mathcal{M}_{1,3}$0

equipped with integrable connection and Hodge filtration. On the Betti side, the standard frame is given by homogeneous polynomials $\mathcal{M}_{1,3}$1 in variables $\mathcal{M}_{1,3}$2, with right action

$\mathcal{M}_{1,3}$3

In level $\mathcal{M}_{1,3}$4, the motive attached to weight $\mathcal{M}_{1,3}$5 cusp forms is realized by the Consani–Faber construction

$\mathcal{M}_{1,3}$6

whose realization is

$\mathcal{M}_{1,3}$7

Here $\mathcal{M}_{1,3}$8 denotes the alternating projector for the $\mathcal{M}_{1,3}$9-action on $(M,N)$0.

The biextension motive is built from marked points. For $(M,N)$1, with $(M,N)$2 containing $(M,N)$3 and $(M,N)$4, and with $(M,N)$5 the forgetful map, one defines

$(M,N)$6

where $(M,N)$7 and $(M,N)$8 are finite unions of residue gerbes, and in the basic case $(M,N)$9, $P$0. Its realization is concentrated in degree $P$1 and identified with

$P$2

This is the mixed modular object from which the biextension structure is extracted (Brown et al., 6 Aug 2025).

3. Exact biextension structures in realizations

The modular biextension is expressed by two compatible exact sequences in $P$3: $P$4

$P$5

where

$P$6

The groups $P$7 and $P$8 are stabilizers of the marked points. These sequences realize $P$9 as a biextension of

$E$0

by the cusp-form motive $E$1.

This formulation is significant because it isolates three arithmetic layers in a single mixed object. The pure modular part $E$2 records cusp-form cohomology. The two outer terms encode symmetric powers of elliptic motives specialized at $E$3 and $E$4. The compatibility of the two extension structures is what permits the arithmetic of Green’s functions to be reinterpreted in motivic terms.

In the CM case, the polarization and the endomorphism structure of $E$5 imply that for $E$6 even, the objects

$E$7

contain Tate summands $E$8 and $E$9. This yields a subquotient

$$0 \to P \to E \to M \to 0, \qquad 0 \to M \to E \to N \to 0,$$0

called the Gross–Zagier extension after Hecke projection. The appearance of a two-step Tate extension is precisely what allows a logarithmic single-valued period to emerge (Brown et al., 6 Aug 2025).

4. Matrix-valued higher Green’s functions as single-valued periods

For integers $$0 \to P \to E \to M \to 0, \qquad 0 \to M \to E \to N \to 0,$$1 with $$0 \to P \to E \to M \to 0, \qquad 0 \to M \to E \to N \to 0,$$2, the generalized higher Green’s functions

$$0 \to P \to E \to M \to 0, \qquad 0 \to M \to E \to N \to 0,$$3

are real-analytic on

$$0 \to P \to E \to M \to 0, \qquad 0 \to M \to E \to N \to 0,$$4

and satisfy modular covariance

$$0 \to P \to E \to M \to 0, \qquad 0 \to M \to E \to N \to 0,$$5

They satisfy generalized Laplacian eigenvalue equations

$$0 \to P \to E \to M \to 0, \qquad 0 \to M \to E \to N \to 0,$$6

These entries are assembled into the generating series

$$0 \to P \to E \to M \to 0, \qquad 0 \to M \to E \to N \to 0,$$7

Its differential equations are governed by a vector-valued meromorphic modular form $$0 \to P \to E \to M \to 0, \qquad 0 \to M \to E \to N \to 0,$$8: $$0 \to P \to E \to M \to 0, \qquad 0 \to M \to E \to N \to 0,$$9 with an analogous equation in the $$\operatorname{Biext}(M,N;P)\cong \operatorname{Ext}^1(M\otimes N,P).$$0-variable.

The single-valued map in $$\operatorname{Biext}(M,N;P)\cong \operatorname{Ext}^1(M\otimes N,P).$$1 is

$$\operatorname{Biext}(M,N;P)\cong \operatorname{Ext}^1(M\otimes N,P).$$2

Applied to the biextension $$\operatorname{Biext}(M,N;P)\cong \operatorname{Ext}^1(M\otimes N,P).$$3, its single-valued period matrix contains a block identified with the Green’s matrix $$\operatorname{Biext}(M,N;P)\cong \operatorname{Ext}^1(M\otimes N,P).$$4. In an explicit de Rham basis, when cusp-form contributions are absent, the $$\operatorname{Biext}(M,N;P)\cong \operatorname{Ext}^1(M\otimes N,P).$$5-matrix takes the form

$$\operatorname{Biext}(M,N;P)\cong \operatorname{Ext}^1(M\otimes N,P).$$6

where $$\operatorname{Biext}(M,N;P)\cong \operatorname{Ext}^1(M\otimes N,P).$$7 is the anti-diagonal identity.

For even weight $$\operatorname{Biext}(M,N;P)\cong \operatorname{Ext}^1(M\otimes N,P).$$8, the central entry recovers the classical higher Green’s function: $$\operatorname{Biext}(M,N;P)\cong \operatorname{Ext}^1(M\otimes N,P).$$9 Thus the classical scalar Green’s function is embedded as a distinguished component of a larger matrix-valued period object. This shift from scalar to matrix-valued periods is one of the main conceptual enlargements: the scalar Gross–Zagier quantity becomes one entry in a biextension period matrix (Brown et al., 6 Aug 2025).

5. Meromorphic modular forms, harmonic lifts, and de Rham realizations

The analytic input is built from rational Poincaré series. For $H_K$0 with $H_K$1,

$H_K$2

and

$H_K$3

These series converge normally on compacta and define meromorphic modular forms in $H_K$4 of weight $H_K$5, vanishing at cusps, with poles along $H_K$6 of order at most $H_K$7 and along $H_K$8 of order at most $H_K$9.

For fixed $\mathbf{Q}$0, the vector-valued generating function

$\mathbf{Q}$1

satisfies

$\mathbf{Q}$2

so it is a harmonic Eichler lift of the reproducing kernel $\mathbf{Q}$3.

A harmonic lift of a holomorphic modular form $\mathbf{Q}$4 of weight $\mathbf{Q}$5 is a $\mathbf{Q}$6-equivariant real-analytic vector-valued function $\mathbf{Q}$7 satisfying

$\mathbf{Q}$8

for some holomorphic modular form $\mathbf{Q}$9, the Betti-conjugate of $4$00. Existence and uniqueness hold in the relevant settings, including meromorphic $4$01. The Bol operator identity

$4$02

controls when such lifts are holomorphic.

These analytic constructions feed directly into algebraic de Rham cohomology. For a field of definition $4$03,

$4$04

Moreover, the Hodge filtration is described in terms of pole orders, and the residue exact sequence splits via the classes $4$05. This identifies the de Rham realization of the biextension with explicit meromorphic modular forms, making the period matrix computable within modular function theory (Brown et al., 6 Aug 2025).

6. Gross–Zagier, Hecke projection, and Kummer reduction

The classical conjecture of Gross and Zagier states that for $4$06, a subgroup $4$07 of finite index, and a Hecke operator

$4$08

annihilating cusp forms of weight $4$09, the Hecke-transformed higher Green’s value $4$10 at CM pairs $4$11 is a rational multiple of $4$12 for an algebraic number $4$13.

In the motivic framework, if $4$14 acts by zero on $4$15, then the biextension admits a Hecke-projected subquotient

$4$16

whose realization is the exact sequence

$4$17

When $4$18 and $4$19 are CM, one extracts the Gross–Zagier extension

$4$20

Its single-valued period is proportional to $4$21.

The decisive motivic input is the standard expectation that in $4$22, every extension of $4$23 by $4$24 is Kummer, with

$4$25

Under the paper’s conjectural hypothesis on Hecke annihilation of $4$26, $4$27 is Kummer, and its single-valued period is $4$28 for some $4$29. Therefore $4$30 is proportional to $4$31.

A special case is proved unconditionally in level $4$32, weight $4$33. The motive $4$34 is shown to be mixed Tate, hence $4$35. Consequently the Gross–Zagier extension is Kummer and the conjecture follows in this case. This replaces analytic arguments with a geometric proof based on the structure of the moduli stack of elliptic curves with three marked points (Brown et al., 6 Aug 2025).

7. Broader arithmetic significance and related biextension theories

The modular construction suggests a broader relation between single-valued periods of biextensions and special values of $4$36-functions. Beyond CM points, Beilinson-type conjectures predict that matrix entries of the Green’s matrix should be related to special values of $4$37-functions of symmetric powers of elliptic curves and Rankin–Selberg convolutions. The paper gives concrete examples of this expectation: for $4$38 and non-CM $4$39, the four entries of $4$40 are expected to relate to

$4$41

while for $4$42 even with $4$43 CM and $4$44 arbitrary over a quadratic imaginary field $4$45, the central column is expected to relate to

$4$46

This suggests that biextensions of modular motives form a natural framework for organizing regulator-type quantities attached to noncritical and derivative $4$47-values.

A related but distinct development appears in the study of Calabi–Yau motives. For rank-$4$48, weight-$4$49 Calabi–Yau motives of analytic rank $4$50, biextension period matrices arise from variations of mixed Hodge structure of type

$4$51

and their minors are compared numerically with $4$52. In modular cases, such as weight-$4$53 $4$54 forms or weight-$4$55 $4$56 paramodular forms, these constructions specialize to biextensions of modular motives, with the period-normalized minors supplying the archimedean regulator term in Beilinson/Bloch–Kato-type formulas (Golyshev, 2023).

The comparison clarifies two complementary uses of biextension technology. In the modular-curve setting, biextensions arise from relative cohomology of the universal elliptic curve with marked points, and their single-valued periods are matrix-valued higher Green’s functions. In the Calabi–Yau setting, biextension period matrices are extracted from Picard–Fuchs systems and interpreted as regulators for rank-$4$57 motives. A plausible implication is that “biextensions of modular motives” should be understood not as a narrowly delimited construction, but as part of a broader motivic mechanism linking mixed extensions, explicit period matrices, and special $4$58-value phenomena across several automorphic and geometric settings (Brown et al., 6 Aug 2025).