Blasius–Deligne Conjecture

• The Blasius–Deligne Conjecture is a framework linking algebraicity of critical motivic and automorphic L-values with period comparisons between Betti and de Rham realizations.
• It refines algebraic predictions by incorporating explicit normalizations using Deligne periods, CM periods, Gauss sums, Petersson norms, and Betti–Whittaker periods.
• The conjecture underpins modern research on symmetric power L-functions, tensor-product formulas, and reciprocity laws connecting motives with automorphic forms.

The Blasius–Deligne conjecture is a family of algebraicity and Galois-equivariance statements for critical values of motivic and automorphic $L$-functions. In Deligne’s formulation, the normalized critical values of a pure motive are controlled by periods $c^{\pm}(M)$ obtained from the comparison between Betti and de Rham realizations. In Blasius-style refinements, especially for Hecke characters, tensor products, and automorphic representations, the normalization becomes more explicit through CM periods, Gauss sums, Petersson norms, Betti–Whittaker periods, or Shalika periods. The modern literature treats this as a coherent program linking Hodge theory, automorphic cohomology, and special values: rank-one CM formulas and their higher-rank analogues, tensor product conjectures for modular forms, and automorphic reciprocity laws all fit within this framework (Lin, 2016).

1. Motivic formulation and Deligne periods

Deligne’s conjecture begins with a pure motive $M$ over $\mathbf Q$, equipped with Betti and de Rham realizations, an archimedean Frobenius action on Betti cohomology, and periods $c^{\pm}(M)$ defined from the comparison isomorphism. For a pure, regular motive $M^\#$ over $\mathbf Q$ with no $(p,p)$-classes, Deligne predicts that for each critical integer $m$,

$$L(m,M^\#)\ \sim_E\ (2\pi i)^{d^\epsilon m}\, c^\epsilon(M^\#),\qquad \epsilon=\mathrm{sign}((-1)^m),$$

where $c^{\pm}(M)$0 means $c^{\pm}(M)$1 and criticality is defined by the absence of poles in both the archimedean factor of $c^{\pm}(M)$2 and the corresponding dual factor at $c^{\pm}(M)$3 (Lin, 2016).

For modular forms, this becomes a concrete statement about motives attached to normalized holomorphic newforms. If $c^{\pm}(M)$4 is a newform of weight $c^{\pm}(M)$5, Deligne–Scholl attach a rank-two motive $c^{\pm}(M)$6, and Deligne computed explicit formulas for $c^{\pm}(M)$7 in terms of $c^{\pm}(M)$8 and $c^{\pm}(M)$9. In particular, the symmetric-power conjecture specializes to algebraicity statements for

$M$0

with the sign $M$1 determined by parity and the precise exponent prescribed by Deligne’s period formulas (Chen, 2022).

Blasius’s tensor-product refinement places the same phenomenon in a higher-rank setting. For normalized holomorphic newforms $M$2 with associated motives $M$3, the conjectural tensor motive $M$4 should satisfy an explicit algebraicity statement at critical points, refined by a period factor $M$5 built from powers of $M$6, Gauss sums of the nebentypoi, and Petersson norms. This is the version that governs triple products and more general tensor-product $M$7-values of modular forms (Chen, 2022).

2. Hecke characters, CM periods, and the reflex-motive formulation

The rank-one case is the classical entry point for the Blasius–Deligne picture. For an algebraic Hecke character $M$8 of a quadratic imaginary field $M$9, Blasius and Deligne predict, and in the CM case prove, that at critical integers $\mathbf Q$0,

$\mathbf Q$1

where $\mathbf Q$2 and $\mathbf Q$3 are CM periods attached to $\mathbf Q$4 and its conjugate. In the normalization used in the automorphic literature, this is compatible with $\mathbf Q$5, and the CM periods are rank-one comparison determinants on distinguished Hodge lines (Lin, 2016).

A more general rank-one formulation applies to algebraic Hecke characters of totally imaginary fields. If $\mathbf Q$6 is a critical algebraic Hecke character of a number field $\mathbf Q$7 with values in $\mathbf Q$8, the period version of Deligne’s conjecture takes the form

$\mathbf Q$9

with $c^{\pm}(M)$0 determined by the critical decomposition of the infinity type. The case $c^{\pm}(M)$1 is the core statement, and the other critical integers follow by twisting (Kufner, 2024).

A major advance is the proof of this conjecture for arbitrary critical algebraic Hecke characters of totally imaginary fields that contain a CM subfield. The key new insight is that the Eisenstein–Kronecker classes of Kings–Sprang can be regarded as de Rham classes of Blasius’ reflex motive. More precisely, the proof constructs a de Rham class $c^{\pm}(M)$2 in the reflex motive $c^{\pm}(M)$3 with

$c^{\pm}(M)$4

while Blasius’s reflex-period relation gives

$c^{\pm}(M)$5

Comparing coefficients yields $c^{\pm}(M)$6, and hence the full critical-value formula by twisting (Kufner, 2024).

3. Automorphic reformulations over CM fields

The automorphic variant replaces motivic periods by periods extracted from cohomology of unitary Shimura varieties. For a cohomological, regular algebraic, conjugate self-dual cuspidal representation $c^{\pm}(M)$7 of $c^{\pm}(M)$8, with $c^{\pm}(M)$9 quadratic imaginary, automorphic periods $M^\#$0 are defined by descending $M^\#$1 to suitable unitary groups, choosing a holomorphic discrete series contribution to coherent cohomology, and taking a normalized Petersson pairing of a rational cohomology class. The resulting quantity $M^\#$2 is well-defined up to $M^\#$3 and depends only on the infinity type of $M^\#$4 and the index $M^\#$5 (Lin, 2016).

The corresponding automorphic conjecture expresses critical values of Rankin–Selberg $M^\#$6-functions in terms of these periods and split indices. If $M^\#$7 and $M^\#$8 are regular, cohomological, conjugate self-dual cuspidal representations of $M^\#$9 and $\mathbf Q$0, then for critical $\mathbf Q$1,

$\mathbf Q$2

The exponents $\mathbf Q$3 are split indices determined solely by the infinity types (Lin, 2016).

This formulation is designed to match the motivic one. Under the expected motivic–automorphic correspondence, one has

$\mathbf Q$4

so the automorphic conjecture is the translation of Deligne’s period formula into automorphic language. In rank one, the automorphic periods reduce to CM periods, and the general statement recovers the quadratic-imaginary formula $\mathbf Q$5 (Lin, 2016).

4. Ratio theorems, Betti–Whittaker periods, and the modern method

A decisive methodological development is the “ratio of special values” strategy for Rankin–Selberg $\mathbf Q$6-functions. For regular algebraic automorphic representations $\mathbf Q$7 of $\mathbf Q$8 and $\mathbf Q$9 of $(p,p)$0, under cohomological, essentially tempered, regularity, and nonvanishing hypotheses, one proves that at a common non-central critical half-integer $(p,p)$1,

$(p,p)$2

with full $(p,p)$3-equivariance. This is the main global ratio theorem for products of Rankin–Selberg values at a fixed critical point (Chen, 2022).

The proof factors each critical value as

$(p,p)$4

Here $(p,p)$5 is a nonzero archimedean constant defined by pairing canonical generators in bottom-degree relative Lie algebra cohomology, while $(p,p)$6 and $(p,p)$7 are algebraic quantities built from Betti–Whittaker periods, Gauss sums, and auxiliary completed Rankin–Selberg values. Passing to the fourfold ratio cancels the archimedean factor and isolates algebraic, Galois-equivariant pieces (Chen, 2022).

Betti–Whittaker periods are the crucial rationality invariants in this argument. For a regular algebraic $(p,p)$8, they are defined by comparing a $(p,p)$9-rational structure on the Whittaker model with a $m$0-rational structure on bottom-degree cohomology. Their construction is $m$1-equivariant, and for Eisenstein representations they are defined multiplicatively in terms of inducing data and $m$2-values at $m$3. The broader method combines Eisenstein cohomology, cohomological realizations of Rankin–Selberg zeta integrals, comparison of rational structures, and a nonvanishing archimedean pairing $m$4 (Chen, 2022).

5. Symmetric powers, tensor products, and modular forms

For symmetric powers of modular forms, the conjecture becomes both explicit and highly structured. In the case of $m$5, if $m$6 is a normalized elliptic newform of weight $m$7, the critical integers are

$m$8

and Chen proves that for every Dirichlet character $m$9 and every such $$L(m,M^\#)\ \sim_E\ (2\pi i)^{d^\epsilon m}\, c^\epsilon(M^\#),\qquad \epsilon=\mathrm{sign}((-1)^m),$$0,

$$L(m,M^\#)\ \sim_E\ (2\pi i)^{d^\epsilon m}\, c^\epsilon(M^\#),\qquad \epsilon=\mathrm{sign}((-1)^m),$$1

transforms exactly as predicted under $$L(m,M^\#)\ \sim_E\ (2\pi i)^{d^\epsilon m}\, c^\epsilon(M^\#),\qquad \epsilon=\mathrm{sign}((-1)^m),$$2. In the same normalization,

$$L(m,M^\#)\ \sim_E\ (2\pi i)^{d^\epsilon m}\, c^\epsilon(M^\#),\qquad \epsilon=\mathrm{sign}((-1)^m),$$3

and the proof also yields period relations between motivic periods of $$L(m,M^\#)\ \sim_E\ (2\pi i)^{d^\epsilon m}\, c^\epsilon(M^\#),\qquad \epsilon=\mathrm{sign}((-1)^m),$$4 and Betti–Whittaker periods of $$L(m,M^\#)\ \sim_E\ (2\pi i)^{d^\epsilon m}\, c^\epsilon(M^\#),\qquad \epsilon=\mathrm{sign}((-1)^m),$$5 on $$L(m,M^\#)\ \sim_E\ (2\pi i)^{d^\epsilon m}\, c^\epsilon(M^\#),\qquad \epsilon=\mathrm{sign}((-1)^m),$$6 (Chen, 2021).

The later general theorem proves Deligne’s conjecture for all symmetric power $$L(m,M^\#)\ \sim_E\ (2\pi i)^{d^\epsilon m}\, c^\epsilon(M^\#),\qquad \epsilon=\mathrm{sign}((-1)^m),$$7-functions of holomorphic modular forms of weight $$L(m,M^\#)\ \sim_E\ (2\pi i)^{d^\epsilon m}\, c^\epsilon(M^\#),\qquad \epsilon=\mathrm{sign}((-1)^m),$$8. If $$L(m,M^\#)\ \sim_E\ (2\pi i)^{d^\epsilon m}\, c^\epsilon(M^\#),\qquad \epsilon=\mathrm{sign}((-1)^m),$$9 is a holomorphic newform and $c^{\pm}(M)$00 is the Newton–Thorne lift to $c^{\pm}(M)$01, then for every critical integer $c^{\pm}(M)$02,

$c^{\pm}(M)$03

is algebraic with the expected $c^{\pm}(M)$04-equivariance, and the periods $c^{\pm}(M)$05 agree with Deligne’s explicit formulas in terms of $c^{\pm}(M)$06 and $c^{\pm}(M)$07 (Chen, 2022).

The same framework proves new cases of Blasius’s tensor-product conjecture. For regular algebraic cuspidal representations $c^{\pm}(M)$08 of $c^{\pm}(M)$09 attached to newforms $c^{\pm}(M)$10, and under explicit regularity and automorphy conditions, one obtains the predicted algebraicity for critical values of

$c^{\pm}(M)$11

with normalization given by powers of $c^{\pm}(M)$12, Gauss sums $c^{\pm}(M)$13, and Petersson norms $c^{\pm}(M)$14. This includes many new unbalanced triple-product cases and higher tensor products under level-one or functoriality hypotheses (Chen, 2022).

6. Symplectic type on $c^{\pm}(M)$15 and unconditional reciprocity

A distinct but closely related branch of the program concerns standard $c^{\pm}(M)$16-functions of symplectic type for $c^{\pm}(M)$17. A regular algebraic cuspidal cohomological representation $c^{\pm}(M)$18 of $c^{\pm}(M)$19 is said to be of symplectic type if there exists a Hecke character $c^{\pm}(M)$20 such that

$c^{\pm}(M)$21

has a pole at $c^{\pm}(M)$22. This is equivalent to $c^{\pm}(M)$23 being essentially self-dual with symplectic sign $c^{\pm}(M)$24, to the existence of a global Shalika model with respect to $c^{\pm}(M)$25, and to the Arthur parameter factoring through $c^{\pm}(M)$26 (Jiang et al., 30 Aug 2025).

In this setting, critical points are encoded by balanced algebraic twists. Writing the infinity type of an algebraic Hecke character as $c^{\pm}(M)$27, with $c^{\pm}(M)$28 quadratic at infinity, one says that $c^{\pm}(M)$29 is $c^{\pm}(M)$30-balanced if a certain archimedean Hom-space for $c^{\pm}(M)$31 is nonzero. Then

$c^{\pm}(M)$32

Thus the critical set is a translate of half-integers, and balancing moves arbitrary critical points to the central point $c^{\pm}(M)$33 (Jiang et al., 30 Aug 2025).

The main theorem gives an unconditional reciprocity identity at the central point, and hence at all critical points by varying $c^{\pm}(M)$34. For every $c^{\pm}(M)$35,

$c^{\pm}(M)$36

and the normalized value lies in $c^{\pm}(M)$37. Here $c^{\pm}(M)$38 encodes the archimedean $c^{\pm}(M)$39-power predicted by Deligne, $c^{\pm}(M)$40 is the $c^{\pm}(M)$41-th power of the global Gauss sum, and $c^{\pm}(M)$42 is a canonically normalized family of Shalika periods. The proof is unconditional and rests on a complete archimedean local theory for Jacquet–Shalika and Friedberg–Jacquet integrals, refined archimedean period relations via Zuckerman translation functors, and global modular symbols (Jiang et al., 30 Aug 2025).

7. Scope, normalizations, and open directions

The conjecture is best viewed as a period-normalization principle rather than a single formula with one universal notation. In the literature surveyed here, “Blasius–Deligne conjecture” may refer to the rank-one CM-period statement for Hecke characters, to tensor-product formulas for modular forms involving Gauss sums and Petersson norms, or to higher-rank automorphic reciprocity laws phrased in terms of automorphic periods, Betti–Whittaker periods, or Shalika periods. This suggests that the term designates a tightly connected family of statements organized by the same period philosophy.

A persistent structural feature is that periods depend on choices—CM types, embeddings, rational bases, coherent cohomology classes, Whittaker generators, or auxiliary characters—but only up to algebraic factors in the relevant rationality fields. The equivalence relations $c^{\pm}(M)$43 and $c^{\pm}(M)$44, together with explicit $c^{\pm}(M)$45-equivariance, are designed to absorb precisely these ambiguities. In the automorphic setting, this is reflected in the definition of $c^{\pm}(M)$46, the Betti–Whittaker periods $c^{\pm}(M)$47, and the Shalika periods $c^{\pm}(M)$48 (Lin, 2016).

The currently proved cases are extensive but remain conditional in some directions. The symmetric-power and tensor-product results on $c^{\pm}(M)$49 require cohomologicality, regularity, explicit weight bounds, and functoriality hypotheses; the automorphic variant over quadratic imaginary fields still leaves open arbitrary $c^{\pm}(M)$50 without gap or parity restrictions; and the motivic isomorphism

$c^{\pm}(M)$51

up to Tate twist is still presented as conjectural input in the automorphic theory over CM fields. By contrast, the Hecke-character theorem beyond CM fields and the symplectic-type $c^{\pm}(M)$52 standard-$c^{\pm}(M)$53 theorem are unconditional within their stated hypotheses (Kufner, 2024).

Taken together, these developments show that large parts of the Blasius–Deligne program are now theorem rather than conjecture: Deligne’s conjecture for all symmetric power $c^{\pm}(M)$54-functions of holomorphic modular forms of weight $c^{\pm}(M)$55, many new cases of Blasius’s tensor-product conjecture, Deligne’s conjecture for arbitrary algebraic Hecke characters of the stated type, and the full symplectic-type standard-$c^{\pm}(M)$56 case on $c^{\pm}(M)$57 have all been established. The remaining open problems are concentrated in the general higher-rank motivic–automorphic dictionary and in removing residual regularity, parity, and descent hypotheses (Chen, 2022).