Weil's closing question of Chapter II (1964): a proof of the general reciprocity law of class field theory analogous to the metaplectic proof of quadratic reciprocity

In his 1964 Acta Mathematica memoir 'Sur certains groupes d'opérateurs unitaires', André Weil derived quadratic (Hilbert) reciprocity from the unitary representation theory of the Heisenberg group: to each nondegenerate quadratic form $f$ over a global field he attached local gamma factors $\gamma_v(f)$ (normalized Gauss sums), proved the product formula $\prod_v \gamma_v(f)=1$ from the Stone–von Neumann theorem and its adèlic summation formula, and identified $\gamma_v$ with Hilbert symbols of the four-variable form $x^2-ay^2-bz^2+abt^2$. He then closed Chapter II with a question, verbatim: 'peut-on donner de la loi générale de réciprocité une démonstration analogue à celle qu'on vient d'exposer pour la loi de réciprocité quadratique ?' — can the general reciprocity law of class field theory (Artin reciprocity) be proved by an analogous route? Later literature extended the metaplectic mechanism to $n$-fold covers of $\mathrm{SL}_2$ and $n$-th power reciprocity (Kubota; Kazhdan–Patterson; cubic theta series and cubic Gauss sums), re-proved quadratic and Hilbert reciprocity by Weil-type methods over finite fields and via Maslov phases (Gurevich–Hadani–Howe; Holland), and Berg framed the analytic approach to general reciprocity laws as an open project; no proof of the general reciprocity law by the metaplectic/theta mechanism is known.

Background

Reciprocity laws relate arithmetic invariants across the completions of a global field. Take a number field $k$ (a finite extension of the rationals; the theory runs identically for function fields of one variable over a finite field, and Weil's memoir works over an arbitrary global field). For each place $v$ of $k$ (an equivalence class of absolute values, corresponding to a completion $k_v$, which is $\mathbb{R}$, $\mathbb{C}$, or a finite extension of the $p$-adic numbers) and units $a,b \in k_v^\times$, the Hilbert symbol $(a,b)_v$ is defined by $(a,b)_v=1$ if the equation $ax^2+by^2=1$ has a solution in $k_v$, and $-1$ otherwise. Hilbert reciprocity is the product formula $\prod_v (a,b)_v = 1$, taken over all places of $k$; for $k=\mathbb{Q}$ it is equivalent to Gauss's quadratic reciprocity law for the Legendre symbol. The general reciprocity law of class field theory — Artin reciprocity — is the corresponding statement for an arbitrary finite abelian Galois extension $L/K$ of global fields. The adèle ring $\mathbb{A}_K$ of $K$ is the restricted topological product of all completions $K_v$ (elements $(x_v)$ with $x_v \in O_v$, the local integer ring, for almost all $v$); the idèle class group is $C_K = \mathbb{A}_K^\times/K^\times$. The Artin symbol attaches to each prime ideal of $K$ unramified in $L$ (one not dividing the relative discriminant of $L/K$) its Frobenius automorphism, i.e. the element of $\mathrm{Gal}(L/K)$ induced on the extension of residue fields, and this extends to a homomorphism from the group of fractional ideals prime to the finite set of ramified primes; class field theory's central theorem (Artin reciprocity) states that the resulting map on $C_K$ is surjective onto $\mathrm{Gal}(L/K)$ with kernel exactly the norm subgroup $N_{L/K} C_L$, so that $\mathrm{Gal}(L/K)$ is identified with a generalized ideal class group of $K$. Tate's essay on Hilbert's ninth problem, 'The general reciprocity law', treats precisely this statement, and this is what Weil's 1964 phrase 'la loi générale de réciprocité' denotes.

Weil's memoir supplies a new mechanism for the quadratic case. To a locally compact abelian group $G$ with Pontryagin dual $\hat G$ (the group of continuous homomorphisms from $G$ to the circle group, itself locally compact and abelian) one attaches the Heisenberg group $A(G)$, the central extension of $G \times \hat G$ by the circle group $\mathbf T$ whose commutator pairing is the canonical one; the Stone–von Neumann theorem states that, up to equivalence, $A(G)$ has a unique irreducible unitary representation with a given nontrivial central character (a homomorphism $\mathbf T \to \mathbf T$ that is not identically $1$). Weil's Theorem 2 is a Poisson-type summation formula for such a representation, and Theorem 5 computes its analytic index (a trace-type invariant). In Chapter II this yields the following: to a nondegenerate quadratic form $f$ (a polynomial with degree-$2$ homogeneous part nondegenerate in the associated bilinear form) on a vector space over a global field, each place $v$ attaches a local gamma factor $\gamma_v(f)$, a complex number of modulus $1$ that is a normalized Gauss sum (a sum of an additive character against a quadratic phase); the adèlic summation formula gives the product formula $\prod_v \gamma_v(f) = 1$. Applied to the norm form of a quaternion algebra (a $4$-dimensional central simple algebra over $k$), this gives Hasse's theorem that a quaternion algebra splits globally if and only if it splits everywhere locally; applied to the four-variable form $x^2 - ay^2 - bz^2 + abt^2$, whose local gamma factor is the Hilbert symbol $(a,b)_v$, it gives Hilbert reciprocity. Weil remarks that the proof 'ne diffère pas substantiellement' from the classical proof of quadratic reciprocity by theta functions (the harmonic-analysis objects generalizing $\sum e^{\pi i n^2 \tau}$) and Gauss sums; his introduction relates it to the treatment in the last chapter of Hecke's classical book on algebraic number fields, and records that P. Cartier had found a direct proof of Theorems 2 and 5.

The same mechanism produces the metaplectic group. The automorphisms of $A(G)$ acting trivially on the center form a group $B_0(G)$, which for $G = X \times X^*$ (with $X$ a finite-dimensional vector space over a local field and $X^*$ its dual) is the symplectic group $\mathrm{Sp}(X)$, the group preserving the standard alternating bilinear form. Uniqueness of the Heisenberg representation makes $\mathrm{Sp}(X)$ act projectively on the representation space, and the associated central extension of $\mathrm{Sp}(X)$ by $\mathbf T$ is the metaplectic group $\mathrm{Mp}(X)$; quotienting by a suitable character gives the familiar double cover of the symplectic group over local and adelic fields. The memoir also treats fields of characteristic $2$, where the quadratic-form theory has special features (additive quadratic forms exist) and Weil introduces variants $\mathrm{Mp}^+(X)$ and $\mathrm{Ps}^+(X)$, proving these extensions are nontrivial. Chapter V develops the framework over algebras with involution, ends with the statement of a Siegel-type formula (an identity generalizing classical results of Siegel) for automorphic functions on classical groups, subject to a rank condition $\dim X_k > 4 \dim Q(X_k/A_k)$, and defers its proof to a sequel, which appeared as Weil's 1965 Acta memoir on the Siegel formula.

After 1964 the metaplectic framework grew in a different direction: Kubota constructed $n$-fold metaplectic covers of $\mathrm{SL}_2$ over global fields and connected them to $n$-th power reciprocity (the reciprocity law for $n$-th power residue symbols); Kazhdan and Patterson's 'Metaplectic forms' developed the analytic theory of automorphic forms on these covers, whose Whittaker coefficients (Fourier coefficients with respect to a nontrivial additive character of the unipotent subgroup) are governed by higher-degree Gauss sums. The cubic ($n=3$) theory is tied to the cubic analogue of the theta series and to cubic Gauss sums, exposed by Deligne in Bourbaki, and continues through recent work of Patterson and of Dunn and Radziwiłł on cubic Gauss sum bias. Gérardin and later Gurevich, Hadani and Howe built the Weil representation over finite fields, the latter using it to prove quadratic reciprocity and the sign of the Gauss sum; Holland gave a metaplectic proof of Hilbert reciprocity in which the local phase is the Kashiwara–Maslov phase (a canonical phase attached to a triple of Lagrangian subspaces, i.e. maximal subspaces on which the alternating form vanishes identically, arising in the analysis of the Heisenberg representation) and the Hilbert symbol is the defect of multiplicativity of these phases. All of these remain within the quadratic or fixed-degree $n$ world: they prove reciprocity for power-residue symbols of a given degree, not the general reciprocity law for arbitrary finite abelian extensions.

Weil closes Chapter II with the question below, immediately after his quadratic reciprocity proof. The question asks whether the general reciprocity law of class field theory can be given a proof built the way Chapter II is built — from local invariants produced by the Heisenberg representation and a global product formula over the places — rather than from the ideal-theoretic or cohomological proofs of class field theory (via Chebotarev's density theorem, an equidistribution statement for how often each Frobenius element occurs among the prime ideals of $K$, or via the group cohomology of Galois groups), which do not use the metaplectic mechanism at all. The theta-function route is known to reach the degree-$2$ case and, via Kubota-type covers, the degree-$n$ power-residue case; the question is whether it reaches the general law.

Problem Statement

Weil's closing question of Chapter II, verbatim (Sur certains groupes d'opérateurs unitaires, Acta Mathematica 111 (1964), end of §30): 'Nous terminerons ce chapitre par une question : peut-on donner de la loi générale de réciprocité une démonstration analogue à celle qu'on vient d'exposer pour la loi de réciprocité quadratique ?' — Can the general reciprocity law be given a proof analogous to the one just exposed for the quadratic reciprocity law?

Here 'the general reciprocity law' is the general reciprocity law of class field theory, that is, Artin reciprocity: for every finite abelian (Galois) extension $L/K$ of global fields, the Artin symbol extends from the unramified primes to a surjective homomorphism from the idèle class group $C_K$ onto $\mathrm{Gal}(L/K)$ whose kernel is the norm subgroup $N_{L/K}C_L$ (equivalently, $\mathrm{Gal}(L/K)$ is canonically identified with a generalized ideal class group of $K$), together with the compatibility of these isomorphisms in towers. The general reciprocity law at this generality covers all finite abelian extensions — both the kernel identification and the surjectivity — the quadratic case, the $n$-th power residue case, and cyclic extensions of a fixed degree being special instances.

'Analogous to that of quadratic reciprocity' refers to the mechanism of Weil's Chapter II: the question is whether the general reciprocity law admits a proof deriving, as Hilbert reciprocity is derived there, from the unitary representation theory of the Heisenberg group — from local invariants attached to second-degree characters via the Stone–von Neumann theorem (gamma factors, equivalently normalized Gauss sums or Weil indices, or their higher-degree and Maslov-phase analogues), combined through a global product or summation formula over the places of the ground field. The problem is to exhibit such a proof: to construct the representation-theoretic local invariants, prove the global formula, and show that the resulting law is Artin reciprocity in full. As the question stands, it admits either a positive answer (such a proof exists) or a demonstration that no proof of this structure can yield the general law; partial reach beyond the quadratic case (for instance to $n$-th power reciprocity via Kubota-type covers) would constitute progress but not a resolution.

References

1. André Weil, "Sur certains groupes d'opérateurs unitaires", Acta Mathematica 111 (1964), 143–211. DOI: 10.1007/BF02391012, https://doi.org/10.1007/BF02391012
2. André Weil, "Sur la formule de Siegel dans la théorie des groupes classiques", Acta Mathematica 113 (1965), 1–87. DOI: 10.1007/BF02391774, https://doi.org/10.1007/BF02391774
3. David A. Kazhdan and S. J. Patterson, "Metaplectic forms", Publications Mathématiques de l'IHÉS 59 (1984), 35–142. DOI: 10.1007/BF02698770, https://doi.org/10.1007/BF02698770
4. Stephen Gelbart, "An elementary introduction to the Langlands program", Bulletin of the American Mathematical Society 10 (1984), 177–219. DOI: 10.1090/S0273-0979-1984-15237-6, https://doi.org/10.1090/S0273-0979-1984-15237-6
5. Shamgar Gurevich, Ronny Hadani, and Roger Howe, "Quadratic reciprocity and the sign of the Gauss sum via the finite Weil representation", International Mathematics Research Notices (2010). DOI: 10.1093/imrn/rnp242; arXiv:0808.2447, https://arxiv.org/abs/0808.2447
6. Michael Berg, "Derived categories and the analytic approach to general reciprocity laws. Part I", International Journal of Mathematics and Mathematical Sciences 2005 (2005), 2133–2158. DOI: 10.1155/IJMMS.2005.2133, https://doi.org/10.1155/IJMMS.2005.2133
7. Michael C. Berg, "Derived Categories and the Analytic Approach to General Reciprocity Laws—Part II", International Journal of Mathematics and Mathematical Sciences 2007 (2007), 1–27. DOI: 10.1155/2007/54217, https://doi.org/10.1155/2007/54217
8. Michael C. Berg, "Derived Categories and the Analytic Approach to General Reciprocity Laws: Part III", International Journal of Mathematics and Mathematical Sciences 2010 (2010), Article 731093. DOI: 10.1155/2010/731093, https://doi.org/10.1155/2010/731093
9. Jonathan Holland, "Reciprocity and the Maslov Phase", arXiv:2604.25288 (2026), https://arxiv.org/abs/2604.25288
10. Samuel J. Patterson, "The Fourier coefficients of metaplectic theta series on GL(2) over rational function fields", arXiv:1411.7594 (2014), https://arxiv.org/abs/1411.7594
11. J. Tate, "Problem 9: The general reciprocity law", Proceedings of Symposia in Pure Mathematics 28.2 (1976), 311–322. DOI: 10.1090/pspum/028.2/0429839, https://doi.org/10.1090/pspum/028.2/0429839
12. Tomio Kubota, "Some results concerning reciprocity law and real analytic automorphic functions", Proceedings of Symposia in Pure Mathematics 20 (1971), 382–395. DOI: 10.1090/pspum/020/0340221, https://doi.org/10.1090/pspum/020/0340221
13. P. Deligne, "Sommes de Gauss cubiques et revêtements de SL(2), d'après S.J. Patterson", in Séminaire Bourbaki 1978/79, exposés 525–542, Lecture Notes in Mathematics 770, pp. 244–277. DOI: 10.1007/bfb0096246, https://doi.org/10.1007/bfb0096246
14. Paul Gérardin, "Weil representations associated to finite fields", Journal of Algebra 46 (1977), 54–101. DOI: 10.1016/0021-8693(77)90394-5, https://doi.org/10.1016/0021-8693(77)90394-5
15. Alexander Dunn and Maksym Radziwiłł, "Bias in cubic Gauss sums: Patterson's conjecture", Annals of Mathematics 200 (2024). DOI: 10.4007/annals.2024.200.3.3, https://doi.org/10.4007/annals.2024.200.3.3
16. S. J. Patterson, "A cubic analogue of the theta series", Journal für die reine und angewandte Mathematik 296 (1977), 125–161. DOI: 10.1515/crll.1977.296.125, https://doi.org/10.1515/crll.1977.296.125