---
title: Higher Hida Theory
url: https://www.emergentmind.com/topics/higher-hida-theory
type: topic
---

# Higher Hida Theory

Higher Hida theory is the extension of classical ordinary Hida theory from degree-zero spaces of modular forms to \(p\)-adic families arising from coherent cohomology in higher degrees, and, in several higher-rank settings, from ordinary or \(\mu\)-ordinary loci on Shimura varieties. In the classical template one projects onto the \(U_p\)-invertible summand by the idempotent \(e_{\mathrm{ord}}=\lim_{n\to\infty} U_p^{n!}\) and organizes the resulting ordinary forms into modules over an Iwasawa weight algebra. Higher Hida theory retains this ordinary-projector philosophy but replaces \(H^0\) by higher coherent cohomology, perfect complexes, or more general ordinary Hecke modules that interpolate across weight space. This framework now appears for Siegel and Hilbert modular varieties, Drinfeld modular curves, and \(\mu\)-ordinary Shimura varieties of Hodge type [1905.08779, 2106.05666, 2507.07423, 2104.11941].

## 1. Classical template and the passage to higher degree

Classical Hida theory, in the number-field case, starts with the modular curve \(X_1(Np^r)\) over \(\mathbf Q_p\), the line bundle \(\omega\), and the Hecke operator \(U_p\) acting on
\[
H^0\bigl(X_1(Np^r),\omega^k\bigr).
\]
The ordinary idempotent is defined by
\[
e_{\mathrm{ord}}=\lim_{n\to\infty} U_p^{n!},
\]
projecting onto the \(U_p\)-invertible summand. The corresponding weight space
\[
W=\operatorname{Spf}\!\bigl(\mathbf Z_p\llbracket \mathbf Z_p^\times\rrbracket\bigr)^{\mathrm{rig}}
\]
parametrizes continuous characters, and a universal character \(\kappa^{\mathrm{un}}\) recovers the classical line bundles \(\omega^k\) after specialization. The control theorem asserts that ordinary parts of \(H^0\) arise by specializing a single finite-flat \(\Lambda\)-algebra, and for \(r\gg 0\) these ordinary parts do not depend on the \(p\)-power level [2507.07423].

Higher Hida theory preserves this architecture while changing the cohomological input. Instead of restricting attention to \(H^0\), it studies ordinary parts of coherent cohomology in positive degrees. On the Drinfeld modular curve, the two relevant coherent groups are
\[
H^0(X,\omega^k), \qquad
H^1\bigl(X,\omega^{1-k}\otimes \omega_D(-2D)\bigr),
\]
which are the only nonzero coherent cohomology groups on the curve [2507.07423]. On the Siegel threefold for \(G=\mathrm{GSp}_4\), one studies
\[
H^i(X_K,[V]), \qquad 0\le i\le 3,
\]
and for regular weights the ordinary part is concentrated in degree \(1\) [1905.08779]. For Hilbert modular varieties of degree \(n=[F:\mathbf Q]\), the relevant coherent cohomology runs through \(0\le q\le n\), producing modules \(M^q\) that interpolate ordinary parts of \(H^q(X,\underline\omega^\kappa)\) [2106.05666].

The defining extension is therefore cohomological rather than merely formal: ordinary \(p\)-adic families are built not only from holomorphic sections but from higher coherent classes, often packaged as perfect complexes over weight space.

## 2. Geometric input: automorphic sheaves, ordinary loci, and Igusa towers

The geometric realization of higher Hida theory depends on an ordinary-type locus and a \(p\)-adic family of sheaves over that locus.

For \(\mathrm{GSp}_4\), let \(X_K\) be a toroidal compactification of the Siegel threefold over \(\mathbf Z_{(p)}\), \(D\) its boundary divisor, and \(A\to X_K\) the universal semi-abelian surface. The vector bundle
\[
\mathcal E = H^1_{\mathrm{dR}}(A/X_K)
\]
has Hodge filtration
\[
0\to \omega_A \to \mathcal E \to \omega_A^{-1}\otimes \nu^{-1}\to 0.
\]
An algebraic representation \(V\) of the Siegel parabolic determines a canonical extension \([V]\) to a locally free sheaf on \(X_K\), and over the \(p\)-ordinary locus \(X_K^{\mathrm{ord}}\) the Igusa tower
\[
\cdots \to \mathrm{IG}(p^m)\to \cdots \to \mathrm{IG}(p)\to X_K^{\mathrm{ord}}
\]
trivializes the conormal bundle of the canonical subgroup. This produces Banach sheaves \(\mathcal F_\kappa\) parametrized by a two-variable weight \(\kappa\in W\simeq \operatorname{Spf}\mathbf Z_p[[X_1,X_2]]\) [1905.08779].

For Hilbert modular varieties in the totally split case, one works with the compactified Shimura variety
\[
X=X_{G,K}\to \operatorname{Spec}(R),
\]
where \(G=\operatorname{Res}_{F/\mathbf Q}\mathrm{GL}_2\), together with the universal semi-abelian scheme \(\mathcal A\to X\) and the Hodge bundle
\[
\omega=e^*\Omega^1_{\mathcal A/X}=\bigoplus_{v\mid p}\omega_v.
\]
A cohomological weight \((\underline k,w)\) defines an automorphic sheaf \(\underline\omega^{(\underline k,w)}\). Over the ordinary locus \(X^{\mathrm{ord}}\), the \(p\)-adic Igusa tower
\[
\pi:\mathrm{IG}\to X^{\mathrm{ord}}
\]
leads to the invertible \(\Lambda\otimes \mathcal O_{X^{\mathrm{ord}}}\)-module
\[
\Omega^\kappa = (\pi_*\mathcal O_{\mathrm{IG}}\otimes \Lambda)^{(\mathbf Z_p^\times)^{n+1},\kappa},
\]
whose specialization recovers \(\underline\omega^{(\underline k,w)}\) on the ordinary locus. The geometry is refined by partial Hasse invariants
\[
h_v\in H^0(X_{\mathbf F_p},\omega_v^{p-1}),
\]
with divisors \(D_v=V(h_v)\), which control the stratification used in the cohomological construction [2106.05666].

For Shimura varieties of Hodge type, the ordinary locus may be empty, and the relevant substitute is the \(\mu\)-ordinary locus
\[
\Sh_K^\mu=\Newton_{\mathrm{max}}\subset \Sh_K\otimes \overline{\mathbf F}_p.
\]
A partial Hasse invariant cuts out precisely the non-\(\mu\)-ordinary locus, and over \(\Sh_K^\mu\) there is a \(G\)-Igusa tower
\[
\Ig_m\to \Sh_K^\mu
\]
with Galois group \(L(\mathbf Z_p)\). The Hodge–Tate map identifies classical sections of automorphic bundles \(\omega^\kappa\) with \(P^{\mathrm{der}}\)-equivariant functions on the Igusa tower, making possible a \(\mu\)-ordinary Hida theory over the weight space
\[
\mathcal W=\operatorname{Spf}\,\mathcal O_{E,(p)}\bigl[\!\bigl[T_P(\mathbf Z_p)\bigr]\!\bigr]
\]
[2104.11941].

In the function-field case of Drinfeld modular curves, the ordinary locus \(\mathfrak X^{\mathrm{ord}}\) similarly supports families of line bundles
\[
\omega^\kappa = \bigl(\mathcal O_{\mathrm{Igusa}}\widehat\otimes R\bigr)^{G,\kappa}
\]
for continuous characters \(\kappa:G=A_\mathfrak p^\times\to R^\times\), interpolating the integral powers \(\omega^k\) [2507.07423].

## 3. Ordinary correspondences and idempotent projectors

The central operator-theoretic input is the construction of ordinary projectors from Hecke correspondences or Frobenius operators that are compact or locally finite on the relevant cohomology.

For \(\mathrm{GSp}_4\), the usual \(U_p\)-operator coming from the Klingen-level correspondence acts compactly on the Banach sheaves \(\mathcal F(R,\kappa)\) and on each cohomology group of the complex computing \(R\Gamma(X_K,\mathcal F(R,\kappa))\). Because \(U_p\) is locally finite on cohomology, the limit
\[
e_{\mathrm{ord}}=\lim_n U_p^{n!}
\]
exists and cuts out the unit-root subspace [1905.08779].

For Hilbert modular varieties, the higher-degree construction requires partial operators. Given a subset \(J\subset \{v\mid p\}\), one forms
\[
T_J=\prod_{v\notin J} U_v \times \prod_{v\in J} F_v
\]
acting on complexes built from twists by the divisors \(D_v\). After applying the idempotent
\[
e(T_J)=\lim_{k\to\infty} T_J^{k!},
\]
the resulting complex is independent of the auxiliary choices and perfect of Tor-amplitude in \([\#J,n]\) [2106.05666].

The Drinfeld case makes the degree dependence especially explicit. Let
\[
X_0(\mathfrak p)\rightrightarrows X
\]
be the Iwahori-at-\(\mathfrak p\) correspondence with universal isogeny \(\pi\). One first defines a naive correspondence
\[
T_\mathfrak p^{\mathrm{naive}}=p_{1,*}\bigl(\pi_k\otimes \mathrm{tr}_{p_1}\bigr)
\]
and then normalizes it by
\[
T_\mathfrak p=\varpi^{-\min(1,k)}T_\mathfrak p^{\mathrm{naive}}.
\]
On the ordinary locus this decomposes as a Frobenius part \(F\), acting on degree-\(1\) cohomology, and a \(U_\mathfrak p\)-operator, acting on degree-\(0\) cohomology. Since these operators are locally finite on
\[
H^0(\mathfrak X^{\mathrm{ord}},\omega^k), \qquad H^1_c(\mathfrak X^{\mathrm{ord}},\omega^k),
\]
one obtains the idempotents
\[
e(U_\mathfrak p)=\lim_{n\to\infty} U_\mathfrak p^{n!}, \qquad
e(F)=\lim_{n\to\infty} F^{n!},
\]
defining the ordinary parts in degrees \(0\) and \(1\) [2507.07423].

For Shimura varieties of Hodge type, the \(\mu\)-ordinary projector is built from partial operators \(U_{p,i}\) indexed by simple positive coroots. Their product
\[
U_p=\prod_i U_{p,i}
\]
is compact and integral, invertible on the \(\mu\)-ordinary locus, and yields
\[
e_{\mathrm{ord}}=\lim_{n\to\infty} U_p^{n!},
\]
which kills the non-\(\mu\)-ordinary part [2104.11941].

Related \(P\)-ordinary theories replace the Borel-ordinary operator by projectors attached to a parabolic subgroup \(P\). In Siegel and unitary settings these take the form
\[
e_P=\lim_{j\to\infty} (U_p^P)^{j!}
\quad\text{or}\quad
e_P=\lim_{n\to\infty}\Bigl(\prod_{w\mid p}\prod_j u_{w,j,\kappa}\Bigr)^{n!},
\]
reflecting a parabolic notion of ordinarity rather than the classical one [1803.10273, 2409.03783].

## 4. Interpolation, control, and algebraic structure

The outcome of the projector formalism is a family of ordinary modules or perfect complexes over weight space, together with specialization isomorphisms recovering classical cohomology.

| Setting | Weight algebra or space | Interpolated ordinary object |
|---|---|---|
| Classical \(\mathrm{GL}_2\) | \(\mathbf Z_p\llbracket \mathbf Z_p^\times\rrbracket\) | \(e_{\mathrm{ord}}H^0(X_1(Np^r),\omega^k)\) |
| \(\mathrm{GSp}_4\) Siegel threefold | \(W\simeq \operatorname{Spf}\mathbf Z_p[[X_1,X_2]]\) | degree-\(1\) ordinary coherent cohomology |
| Hilbert modular variety | \(\Lambda\simeq R[[X_1,\dots,X_{n+1}]]\) | \(M^q\) interpolating \(H^q(X,\underline\omega^\kappa)^{\mathrm{ord}}\) |
| Drinfeld modular curve | \(\Lambda=A_\mathfrak p\llbracket A_\mathfrak p^\times\rrbracket\) | \(M^{\mathrm{ord}}\) in degree \(0\), \(N^{\mathrm{ord}}\) in degree \(1\) |
| Hodge type Shimura variety | \(\mathcal W=\operatorname{Spf}\mathcal O_{E,(p)}[[T_P(\mathbf Z_p)]]\) | \(\mu\)-ordinary \(H^0\)-families |

For \(\mathrm{GSp}_4\), Pilloni’s existence and control theorem states that, for \(\kappa\) sufficiently close to classical regular weights, the ordinary part
\[
M^i(\kappa):=e_{\mathrm{ord}}\,R\Gamma^i(X_K,\mathcal F(R,\kappa))
\]
is concentrated in degree \(i=1\), is a finite projective \(R\)-module, and specializes canonically to
\[
e_{\mathrm{ord}}\,H^1\bigl(X_K\otimes \mathbf Q_p,\omega^{(r_2-r_1,r_2)}(-D)\bigr)
\]
with Hecke equivariance. Equivalently, one obtains a single perfect complex of \(R\)-modules of amplitude \([1,1]\), interpolating the ordinary part of \(H^1\) for regular cohomological weights [1905.08779].

For Hilbert modular varieties, the complexes
\[
M_J(\Omega^\kappa)=\Bigl[\varprojlim_i e(T_J)\,\mathrm{R}\Gamma_i^J(\Omega^\kappa)\Bigr]
\]
lie in \(D^b_{\mathrm{perf}}(\Lambda)\), and their cohomology in degree \(q=\#J\) is denoted \(M^q\). The interpolation theorem gives
\[
M_J(\Omega^\kappa)\otimes_{\Lambda,(\underline k,w)} R
\simeq
e(T_p)\,R\Gamma\bigl(X,\underline\omega^{(\underline k,w)}\bigr),
\]
hence
\[
M^q\otimes_{\Lambda,(\underline k,w)} R
\simeq
H^q\bigl(X,\underline\omega^{(\underline k,w)}\bigr)^{\mathrm{ord}}.
\]
Each \(M^q\) is finitely generated, torsion-free, and \(\Lambda\)-projective of finite rank [2106.05666].

For Drinfeld modular curves, one sets
\[
M=H^0(\mathfrak X^{\mathrm{ord}},\omega^{\kappa^{\mathrm{un}}}),\qquad
N=H^1_c(\mathfrak X^{\mathrm{ord}},\omega^{\kappa^{\mathrm{un}}}),
\]
and defines
\[
M^{\mathrm{ord}}=e(U_\mathfrak p)M,\qquad
N^{\mathrm{ord}}=e(F)N.
\]
These are direct summands finite over the semilocal, non-Noetherian Iwasawa algebra \(\Lambda=A_\mathfrak p\llbracket G\rrbracket\). For \(k\ge 3\),
\[
M^{\mathrm{ord}}\otimes_{\Lambda,f_k} A_\mathfrak p
\cong
e(T_\mathfrak p)H^0(X,\omega^k),
\]
and for \(k\le -1\),
\[
N^{\mathrm{ord}}\otimes_{\Lambda,f_k} A_\mathfrak p
\cong
e(T_\mathfrak p)H^1\bigl(X,\omega^{1-k}\otimes \omega_D(-2D)\bigr),
\]
so both cohomological degrees are interpolated [2507.07423].

For Hodge-type Shimura varieties, Zhang proves that
\[
M:=V_\infty^{\mathrm{ord}}
\]
is finite projective over \(\mathcal O(\mathcal W)^\circ\), and for an arithmetic weight \(\chi\),
\[
M\otimes_{\mathcal O(\mathcal W)^\circ,\chi} L
\simeq
e_{\mathrm{ord}}\,H^0(\Sh_K,\omega^\chi).
\]
For sufficiently regular algebraic \(\kappa\), higher cohomology vanishes and the specialization map is an isomorphism [2104.11941].

## 5. Duality, \(\Lambda\)-adic Hodge theory, and arithmetic applications

Higher Hida theory is not only an interpolation statement; it also carries structural dualities and arithmetic output.

In the Drinfeld setting, the family version of Serre duality is explicit. There is a perfect pairing of \(\Lambda\)-modules
\[
M^{\mathrm{ord}}\times N^{\mathrm{ord}}\longrightarrow \Lambda
\]
whose specialization recovers the usual Serre duality pairing between
\[
e(T_\mathfrak p)H^0(X,\omega^k)
\quad\text{and}\quad
e(T_\mathfrak p)H^1\bigl(X,\omega^{1-k}\otimes \omega_D(-2D)\bigr)
\]
[2507.07423].

Cais’s work on the geometry of Hida families provides a \(\Lambda\)-adic Hodge-theoretic underpinning for ordinary families. It constructs \(\Lambda\)-adic Dieudonné and crystalline cohomology, proves \(\Lambda\)-adic comparison isomorphisms with de Rham and étale cohomology, gives a cohomological construction of the family of \((\varphi,\Gamma)\)-modules attached to Hida’s ordinary \(\Lambda\)-adic étale cohomology, supplies a new geometric proof of Hida’s finiteness and control theorems, and establishes compatible \(\Lambda\)-adic duality theorems in Dieudonné, crystalline, and étale realizations [1407.5709].

On the analytic side, the theory has become a tool for constructing \(p\)-adic \(L\)-functions from coherent cohomology. For \(\mathrm{GSp}_4\), Loeffler–Pilloni–Skinner–Zerbes use higher Hida theory to \(p\)-adically interpolate periods of non-holomorphic automorphic forms and to construct \(p\)-adic \(L\)-functions for the degree \(4\) spin \(L\)-function of automorphic representations of \(\mathrm{GSp}_4\) and for the degree \(8\) \(L\)-function of \(\mathrm{GSp}_4\times \mathrm{GL}_2\) [1905.08779]. In the Hilbert case, the higher Hida modules are described as essential in constructing \(p\)-adic \(L\)-functions via coherent cohomology classes, including Asai and triple-product \(L\)-functions, and in formulating Bloch–Kato conjectures in higher rank [2106.05666].

Adjacent ordinary theories lead to related arithmetic applications. Non-cuspidal Hida theory for Siegel modular forms is used to construct improved \(p\)-adic \(L\)-functions and prove a derivative formula at a semi-stable trivial zero, verifying a higher-rank analogue of the Mazur–Tate–Teitelbaum/Greenberg conjecture for standard \(L\)-functions of \(P\)-ordinary Siegel cusp forms [1803.10273]. For unitary groups, \(P\)-ordinary Hida families together with Schneider–Zink types support a doubling-method construction of \(p\)-adic \(L\)-functions interpolating standard \(L\)-values [2409.03783].

The Drinfeld theory also points toward Galois-theoretic applications: each ordinary summand \(M^{\mathrm{ord}}\) and \(N^{\mathrm{ord}}\) is expected to carry a Galois representation valued in \(\Lambda\), interpolating the \(p\)-adic Galois representations associated to classical Drinfeld cuspforms, and local–global compatibility at places \(\mathfrak p\) is expected for these families [2507.07423].

## 6. Variants, scope, and common ambiguities

The current literature shows that “higher Hida theory” is not a single construction with uniform inputs. In one direction, it means interpolation of higher coherent cohomology, as in the Siegel, Hilbert, and Drinfeld settings [1905.08779, 2106.05666, 2507.07423]. In another direction, it denotes higher-rank ordinary theories on more general Shimura varieties, where the cohomological realization may remain in \(H^0\) but the ordinary geometry is replaced by the \(\mu\)-ordinary locus and its associated parabolic \(P_\mu\) [2104.11941]. A further extension replaces Borel ordinarity by \(P\)-ordinarity, as in the Siegel and unitary-group theories attached to a parabolic subgroup \(P\) [1803.10273, 2409.03783].

A common ambiguity concerns the adjective “higher.” In the coherent-cohomological literature it refers to higher degrees of automorphic cohomology; in higher-rank ordinary theories it may refer instead to the group or the parabolic structure. Another nearby but distinct theme is higher codimension Iwasawa theory for tensor products of Hida families. Lei–Palvannan study Galois representations such as \(\rho_{4,3}\) and \(\rho_{8,4}\), two \(p\)-adic \(L\)-functions arising from distinct Panchishkin conditions, and codimension-two cycles
\[
c_2(R/I)
\]
attached to the ideal generated by these \(p\)-adic \(L\)-functions. This is built from Hida families, but its central objects are pseudo-null modules and height-two cycles rather than higher coherent cohomology [1901.09301].

The term also intersects with local characterization results for Hida families. For genus-two Siegel modular forms, Hida families arising from stable Yoshida lifts can be characterized by density of de Rham specializations at singular weights \((k,2)\) and by local decomposability at \(p\) of the associated \(\Lambda\)-adic Galois representation, under pseudo-nullity assumptions on stricter Selmer groups [2602.20737]. These results are not a construction of higher coherent cohomology, but they illustrate how ordinary \(\Lambda\)-adic families interact with local \(p\)-adic Hodge-theoretic conditions.

Several open directions are explicit in the existing works. In the Drinfeld setting, analogues for higher-rank Drinfeld and more general function-field Shimura varieties are expected to follow similar lines [2507.07423]. In the tensor-product setting, the unbalanced \(4\)-variable \(L_p^{\mathrm{unb}}\) remain conjectural [1901.09301]. For \(P\)-ordinary unitary groups, the big \(P\)-ordinary Hecke algebra is conjecturally finite-flat over \(O_\pi[[Z_P^\circ]]\), while the Borel case is known by Hida [2409.03783].

Taken together, these developments show that higher Hida theory has become a cohomological and geometric framework for packaging ordinary automorphic data into \(p\)-adic families over weight space, with control theorems, dualities, and arithmetic applications that extend far beyond the original degree-zero theory.

Source: https://www.emergentmind.com/topics/higher-hida-theory