---
title: Shalika's Germ Expansion in Orbital Integrals
url: https://www.emergentmind.com/topics/shalika-s-germ-expansion
type: topic
---

# Shalika's Germ Expansion in Orbital Integrals

Shalika’s germ expansion is a local asymptotic formula for orbital integrals. In its classical group-theoretic form, for a regular semisimple element \(t\) approaching the identity or, more generally, a central element \(z\in Z(F)\), the orbital integral \(O_G(t,f)\) is expressed as a linear combination of orbital integrals at unipotent elements, with coefficients called germs. In the Lie algebra formulation, normalized orbital integrals near the nilpotent cone are expanded in a basis indexed by nilpotent orbits. Recent work has emphasized three complementary aspects of the subject: explicit recursive computation through twisted Levi descent, combinatorial formulas for tamely ramified elements in \(GL_n\), and an endoscopic replacement for the classical expansion in \(SL(2)\) when the usual unipotent indexing set ceases to be finite [1512.00445] [2209.02509] [2507.05003].

## 1. Classical form of the expansion

For \(G=SL(2)\) over a non-archimedean local field \(F\), the group version near a central element \(z\) is written, for \(f\in C_c(G(F))\) and regular semisimple \(t\) close to \(z\), as
\[
O_G(t,f)= \Gamma^G_1(t)f(z)+\sum_{u\in U} \Gamma^G_{u}(t)O_G(zu,f),
\]
where \(U\) is the set of nontrivial unipotent conjugacy classes in \(G(F)\), \(\Gamma^G_1(t)\) is the germ attached to the identity class, and \(\Gamma^G_u(t)\) is the germ attached to the unipotent class \(u\) [2507.05003]. In this formulation, the asymptotics of semisimple orbital integrals near the identity are controlled by unipotent orbital integrals.

The Lie algebra version, as summarized in the setting of DeBacker’s enhanced theorem, takes the form
\[
I^G(\gamma,f)=\sum_{O\in \mathcal O^G(0)}\Gamma_O(\gamma)\, I^G(O,f),
\]
for \(\gamma\in \mathfrak g\) and suitable test functions \(f\in C_c^\infty(\mathfrak g)\), with \(\mathcal O^G(0)\) the set of nilpotent orbits and \(\Gamma_O(\gamma)\) the normalized Shalika germs [1512.00445]. The same organizational principle appears in the Iwahori-biinvariant Hecke-algebra setting for \(GL_n\), where for a regular semisimple \(\gamma\in G(O)\),
\[
I_\gamma(f)=\sum_{\lambda} \Gamma_\lambda(\gamma)\, I_\lambda(f), \qquad f\in C_c^\infty(G(F)/I),
\]
with partitions \(\lambda\vdash n\) indexing unipotent or nilpotent orbits [2209.02509].

These formulas exhibit the central feature of Shalika germs: they are the coefficients that connect regular semisimple orbital integrals to nilpotent or unipotent orbital integrals. In the classical setting, the relevant indexing set must be finite if the resulting asymptotic formula is to remain a finite sum [2507.05003].

## 2. Indexing by nilpotent and unipotent orbits

The indexing of the expansion depends on whether one is in the Lie algebra or group setting. In the Lie algebra formulation, the coefficients are indexed by nilpotent orbits \(O\in \mathcal O^G(0)\), and the orbital integrals \(I^G(O,f)\) provide the nilpotent basis against which the regular semisimple orbital integral is expanded [1512.00445]. In the \(GL_n\) group formulation studied for tamely ramified elements, the index set is the set of partitions \(\lambda\vdash n\), which parameterizes unipotent or nilpotent orbits of \(GL_n\) [2209.02509].

For \(SL(2)\), the unipotent indexing set has a particularly explicit description. The paper on \(SL(2)\) states that the determinant induces an isomorphism
\[
G(F)\backslash GL(2,F)/G(F)\simeq Q_F,\qquad Q_F:=(F^\times)^2\backslash F^\times,
\]
and that the set \(U\) of conjugacy classes of nontrivial unipotent elements in \(G(F)\) is in noncanonical bijection with \(Q_F\) [2507.05003]. Concretely, for
\[
u_0=\begin{pmatrix}1&1\\0&1\end{pmatrix},
\]
the map
\[
\nu:U(F)\to F,\qquad  
\begin{pmatrix}1&\eta\\0&1\end{pmatrix}\mapsto \eta
\]
induces a bijection between rational conjugacy classes of nontrivial unipotent elements and \(Q_F\), with \(u_0\) corresponding to \(1\in Q_F\) [2507.05003].

This explicit parametrization clarifies why the size of the indexing set is decisive. The general expansion is structurally simple only when the space of rational unipotent classes remains finite. The \(SL(2)\) case shows that this is a substantive hypothesis rather than a mere technical convenience [2507.05003].

## 3. Recursive computation and twisted Levi descent

An important modern development is the conversion of Shalika germs from formal coefficients into recursively computable objects. The paper on the inductive structure of Shalika germs begins with a non-nilpotent \(\gamma\in \mathfrak g\) and a depth decomposition
\[
\gamma=\gamma_0+\gamma',
\]
where \(\gamma_0\) is good of depth \(d=\operatorname{depth}(\gamma)\), \([\gamma,\gamma_0]=0\), and \(\gamma'\) has strictly larger depth [1512.00445]. One then sets
\[
G' = Z_G(\gamma_0),
\]
which is a twisted Levi subgroup. This is the basic reduction step.

The recursive engine is an orbital-integral identity that reduces computation on \(G\) to computation on \(G'\). For special test functions \(f_i\),
\[
I^G(\gamma,f_i) = \sum_{w\in W_{x_i}}
\frac{|D^G(\gamma)|^{1/2}\,|G_{x_i,0+}|}
{|D^{G'}(\gamma')|^{1/2}\,|G'_{w^{-1}x_i}|}
\, I^{G'}(\gamma',\widetilde f'_{i,w,\gamma_0}),
\]
and the Shalika germ expansion in \(G'\) gives
\[
I^{G'}(\gamma',\widetilde f'_{i,w,\gamma_0})
= \sum_{O'\in\mathcal O^{G'}(0)} \Gamma_{O'}(\gamma')\,I^{G'}(O',\widetilde f'_{i,w,\gamma_0}) .
\]
Thus germs for \(G\) are computed from germs for \(G'\), together with finite combinatorial data and point counts on explicit varieties over the residue field [1512.00445].

The same paper constructs a finite family of test functions \(f_i\) supported on Moy–Prasad pieces so that the matrix \(\bigl(I^G(O,f_i)\bigr)_{O,i}\) is block upper triangular with nonzero diagonal, hence invertible by induction on orbit dimension [1512.00445]. It also identifies the geometric objects appearing in the algorithm as quasi-finite covers of Hessenberg varieties. A basic example is
\[
\hat H_e(\gamma) = \{g\in G^{(0)}/E^o \mid \operatorname{Ad}(g^{-1})(\gamma)\in e+g_{\le 1}^{(1)}\},
\]
and the paper states that all varieties occurring in the algorithm are of this form and that these varieties are smooth [1512.00445].

A parallel affine-Springer-fiber recursion is established through schemes \(X_{x,d,e}\) and \(Y_{x,d,e}\), with vanishing, depth-lowering, and stratification statements that mirror the Shalika-germ recursion [1512.00445]. This places the harmonic analysis and the geometry in a common inductive framework.

## 4. Tamely ramified \(GL_n\): symmetric functions and combinatorics

For tamely ramified regular semisimple elements in \(GL_n\), the recent literature gives an explicit combinatorial realization of Shalika germs. The starting point is a Puiseux-type expansion
\[
\gamma=\sum_{r\in \frac1n \mathbb{Z}_{\ge 0}} a_r u^{nr},\qquad a_r\in O^\times,
\]
from which one extracts the finite set of root valuations
\[
RV(\gamma)=\{\,\operatorname{val}_F(\sigma(\gamma)-\gamma)\;|\;\sigma\in\operatorname{Gal}(F^{sep}/F)\,\}.
\]
These determine Puiseux pairs \((m_i,n_i)\), then Newton pairs \((p_i,q_i)\), and for inertially elliptic \(\gamma\) the master function and germ expansion depend only on these Newton pairs [2209.02509].

The central object is a symmetric function \(f\), the master symmetric function. In the totally ramified case it is defined recursively by
\[
f=\varphi_{q_d/p_d}\Big(\cdots \varphi_{q_1/p_1}(e_1)\cdots\Big),
\]
and the main theorem identifies the Shalika germs through
\[
\sum_{\lambda\vdash n}\Gamma_{\lambda^t}(\gamma)\,h_\lambda
= q^{\Xi(\gamma)}\,\omega\!\left(f\big|_{q\mapsto q^{-1}}\right),
\]
where \(\Xi(\gamma)\) is a universal \(q\)-power shift and the transpose \(\lambda\mapsto\lambda^t\) is part of the normalization forced by the relation between unipotent orbital integrals and the Hall inner product [2209.02509].

A conceptual reformulation is provided by the elliptic Hall algebra. For each slope \(m/n\) there is a commutative subalgebra \(E^{m/n}\), and an isomorphism
\[
\widehat\varphi_{m/n}:\Lambda_{q,t}\xrightarrow{\sim} E^{m/n}, \qquad p_k\mapsto P_{km,kn}.
\]
At \(t=1\), the relevant operators degenerate to multiplication operators, and one obtains the combinatorial formula
\[
\varphi_{m/n}(e_k)\Big|_{t=1} = E_{m,n,k}
= \sum_{D\in D_{km,kn}} q^{\operatorname{area}(D)} e_D,
\]
with \(D_{km,kn}\) the set of Dyck paths in the \(km\times kn\) rectangle [2209.02509]. The transition matrices governing one step of the recursion are then expressed through symmetric-group cycle data, Dyck paths, compositions refining partitions, and \(q\)-weights from area and coarea statistics [2209.02509].

This framework also produces orbital-integral formulas. For a partition \(\lambda\vdash n\), if \(P_\lambda\) is the standard parahoric subgroup and \(1_\lambda\) its normalized characteristic function, then
\[
I_\gamma(1_\lambda)=\langle f,h_\lambda\rangle.
\]
When \(F=k((t))\), these orbital integrals count points on affine Springer fibers, implying polynomiality and positivity statements for point counts, and via the Laumon correspondence also for local compactified Jacobians of planar curves [2209.02509]. The same paper further proposes a conjectural interpretation on \(\Hilb^n(\mathbb A^2)\), where a quasi-coherent sheaf \(\mathcal F_\gamma\) is expected to provide a \(t\)-deformation of the germs [2209.02509].

## 5. \(SL(2)\) in arbitrary characteristic

The paper on \(SL(2)\) isolates a basic obstruction to the classical formulation. For
\[
Q_F=(F^\times)^2\backslash F^\times,
\]
the set \(U\) of nontrivial rational unipotent conjugacy classes is compact, finite if \(p\neq 2\), and uncountable if \(p=2\) [2507.05003]. Consequently, for elliptic elements in characteristic \(2\), the classical Shalika expansion
\[
O_G(t,f)= \Gamma^G_1(t)f(z)+\sum_{u\in U} \Gamma^G_u(t)O_G(zu,f)
\]
does not make sense as a finite germ expansion, because the indexing set is uncountable [2507.05003].

The same work introduces \(\kappa\)-orbital integrals using characters \(\kappa\) of \(Q_F\):
\[
O_G^\kappa(x,f)= \int_{Q_F}\kappa(q)O_G(x^q,f)\,dq,
\]
with Fourier inversion
\[
O_G(x,f)=\sum_{\kappa\in \widehat{Q_F}} O_G^\kappa(x,f).
\]
The case \(\kappa=1\) gives the stable orbital integral \(SO_G(x,f)\), while \(\kappa\neq 1\) gives unstable pieces; the paper also notes
\[
O_G(z,f)=SO_G(z,f),\qquad O_G^\kappa(z,f)=0\ \text{for}\ \kappa\neq 1.
\]
When \(p\neq 2\), Fourier inversion on the finite group \(Q_F\) rewrites the classical Shalika expansion in \(\kappa\)-germ form, linking unipotent germs to endoscopy [2507.05003].

The principal replacement valid in all characteristics is an endoscopic expansion. For \(SL(2)\), an endoscopic pair is either \(\{G,1\}\) or \(\{E/F,\kappa\}\), where \(E/F\) is a separable quadratic extension and \(\kappa\) the associated quadratic character. If \(f^E\in C_c(H(F))\) is an endoscopic transfer of \(f\), then
\[
SO_H(\iota(t),f^E)=\Delta^E(t)\,O_G^\kappa(t,f)
\]
for all regular \(t\in T(F)\), and such a transfer always exists [2507.05003]. The main theorem states that for regular \(t\in T(F)\) close to \(z\in Z(F)\),
\[
O_G(t,f)= \sum_{E=\{H,\kappa\}\in \mathrm{ell}} \Phi^E(t),
\]
with \(\Phi^E(t)\) defined in terms of \(\Delta^E(t)\) and stable orbital integrals on the endoscopic group, and that the stable orbital integral on the endoscopic group admits a germ expansion indexed by stable unipotent conjugacy classes:
\[
SO_H(f^E(\iota(t)))= \sum_{u\in SU^H} S^H_{\iota(z)u}(\iota(t))\, SO_H(\iota(z)u,f^E).
\]
For \(p\neq 2\), this endoscopic expansion is equivalent, up to Fourier transform, to the classical Shalika expansion; for \(p=2\), it is genuinely new [2507.05003].

A further simplifying lemma states that for \(SL(2)\) and \(GL(2)\), stable conjugacy coincides with conjugacy under the relevant group [2507.05003]. In this setting, the contrast between ordinary and stable objects is therefore expressed through the germ expansions and transfer relations rather than through a separate stable-conjugacy classification.

## 6. Geometric scope, applications, and structural lessons

The recent literature presents Shalika germs as part of a larger structure linking harmonic analysis, arithmetic geometry, and endoscopy. The inductive framework of twisted Levi descent reduces general germs to smaller groups, finite combinatorial data, and point counts of explicit varieties over the residue field [1512.00445]. The tamely ramified \(GL_n\) theory packages the same information into a master symmetric function controlled by Newton pairs, slope plethysms, and the degeneration of the elliptic Hall algebra at \(t=1\) [2209.02509]. The \(SL(2)\) theory shows that endoscopy can supply the correct local asymptotic expansion even when the classical unipotent index set is not finite [2507.05003].

Several consequences are explicitly recorded. In the \(GL_n\) setting, the formula
\[
I_\gamma(1_\lambda)=\langle f,h_\lambda\rangle
\]
identifies orbital integrals with Hall pairings against the master symmetric function, and when \(F=k((t))\) these orbital integrals count points on affine Springer fibers [2209.02509]. The paper further states polynomiality in \(q\) with nonnegative coefficients for point counts of affine Springer fibers and, via the Laumon correspondence, for local compactified Jacobians of planar curves, and relates the same framework to the ORS conjecture and a virtual version of the Cherednik–Danilenko conjecture [2209.02509]. In the more general inductive setting, the paper proves formal applications including local constancy under sufficiently deep central perturbations, homogeneity near singular semisimple elements, and uniform bounds for certain orbital integrals [1512.00445].

A common misconception is that Shalika’s germ expansion is uniformly a finite sum indexed by unipotent classes. The \(SL(2)\) characteristic-\(2\) case shows that this fails at the level of rational unipotent conjugacy classes, so that the classical formulation for elliptic elements does not make sense there [2507.05003]. Another misconception is that the recent endoscopic expansion is merely a reformulation of the older theory. The paper on \(SL(2)\) states that this is true only when \(p\neq 2\), where equivalence holds up to Fourier transform; in characteristic \(2\), the endoscopic expansion is new [2507.05003].

Taken together, these developments portray Shalika’s germ expansion not as a single formula but as a family of local asymptotic mechanisms. In one regime it is a finite expansion against nilpotent or unipotent orbital integrals; in another it becomes an explicit combinatorial object encoded by symmetric functions; and in the \(SL(2)\) characteristic-\(2\) regime it is replaced by an endoscopic expansion indexed by elliptic endoscopic data rather than by ordinary unipotent classes [1512.00445] [2209.02509] [2507.05003].

Source: https://www.emergentmind.com/topics/shalika-s-germ-expansion