---
title: 'Mirabolic Hecke Algebra: Structure & Dualities'
url: https://www.emergentmind.com/topics/mirabolic-hecke-algebra
type: topic
---

# Mirabolic Hecke Algebra: Structure & Dualities

The mirabolic Hecke algebra is the type \(A\) Hecke-theoretic object obtained by adjoining a vector to the usual double-flag geometry. In the finite-field model, with \(G=GL_d(\mathbb F_q)\), complete flag variety \(\mathcal Y\), and vector space \(V=\mathbb F_q^d\), it is the convolution algebra of \(G\)-invariant functions on \(\mathcal Y\times \mathcal Y\times V\). Algebraically it may be presented by generators \(T_0,T_1,\dots,T_{d-1}\), where \(T_1,\dots,T_{d-1}\) satisfy the ordinary type-\(A\) Hecke relations and \(T_0\) encodes the additional mirabolic datum. In the recent literature the same algebra appears under the notations \(MH\), \(R_n\), and \(\mathscr H_{n,R}(q)\), depending on normalization and base ring [2404.05594], [1310.3878], [2603.00603].

| Notation | Setting | Source |
|---|---|---|
| \(MH\) | generic mirabolic Hecke algebra via convolution on \(\mathcal Y\times\mathcal Y\times V\) | [2404.05594] |
| \(R_n\) | Rosso’s generic mirabolic Hecke algebra | [1310.3878] |
| \(\mathscr H_{n,R}(q)\) | arbitrary-commutative-ring presentation | [2603.00603] |

## 1. Geometric origin and orbit combinatorics

The adjective “mirabolic” refers to the mirabolic subgroup, namely the stabilizer of a nonzero vector. A basic geometric principle recalled in the literature is that, for a \(GL_d(\mathbb C)\)-variety \(X\), \(M\)-orbits on \(X\) correspond bijectively to \(GL_d(\mathbb C)\)-orbits on \(X\times (V\setminus\{0\})\). In the finite-field incarnation used for mirabolic Hecke algebras, this principle becomes a convolution theory on spaces of the form
\[
\mathcal X\times\mathcal X\times V,\qquad
\mathcal Y\times\mathcal Y\times V,\qquad
\mathcal X\times\mathcal Y\times V,
\]
where \(\mathcal X\) is a partial flag variety and \(\mathcal Y\) is the complete flag variety [2404.05594].

For the Hecke side one uses complete flags. The \(G\)-orbits on \(\mathcal Y\times\mathcal Y\times V\) are finite. They are parametrized by pairs \((w,\beta)\), where \(w\in S_d\) and \(\beta\) satisfies a monotonicity condition; equivalently, \(\beta\) may be written as a strictly decreasing sequence
\[
\beta=(j_1,\dots,j_k),\qquad 1\le j_k<\dots<j_1\le d,
\]
such that
\[
w(j_1)<w(j_2)<\dots<w(j_k).
\]
The characteristic function of the orbit corresponding to \((w,\beta)\) is denoted \(T_{w,\beta}\). The elements \(T_{s_i,0}\) attached to simple transpositions and \(T_{\mathrm{id},\{1\}}\) supply the basic generators of the algebra [2404.05594].

The convolution product is defined by summing over an intermediate complete flag and an intermediate vector:
\[
(g*h)(\mathfrak f,\mathfrak f',w)=
\sum_{\mathfrak f''\in\mathcal Y,\;p\in V}
g(\mathfrak f,\mathfrak f'',p)\,h(\mathfrak f'',\mathfrak f',w-p).
\]
This product is associative, and the \(G\)-invariant characteristic functions \(T_{w,\beta}\) form a basis. In this sense the mirabolic Hecke algebra is the direct analogue of the Iwahori–Hecke algebra of type \(A\), except that the geometry has been enlarged from pairs of flags to triples consisting of two flags and a vector [1310.3878].

The extra vector is not a superficial modification. In the non-mirabolic setting orbit data are controlled by relative position of flags alone; in the mirabolic setting the vector produces a second layer of incidence data. On partial-flag varieties this is encoded by decorated matrices, and for \(n>2\) their behavior under convolution is explicitly described as more intricate than in the rank-one case [2404.05594].

## 2. Algebraic presentations

In Solomon’s presentation, the mirabolic Hecke algebra is the associative algebra generated by
\[
T_0,T_1,\dots,T_{n-1}
\]
with relations
\[
T_0^2=(q-2)T_0+(q-1),
\qquad
T_i^2=(q-1)T_i+q\quad (i\ge1),
\]
the ordinary type-\(A\) braid relations for \(T_1,\dots,T_{n-1}\),
\[
T_iT_j=T_jT_i\qquad (|i-j|\ge2),
\]
and the two mixed relations
\[
T_0T_1T_0T_1=(q-1)(T_1T_0T_1+T_1T_0)-T_0T_1T_0,
\]
\[
T_1T_0T_1T_0=(q-1)(T_1T_0T_1+T_0T_1)-T_0T_1T_0.
\]
In the normalization \(q=v^2\), these are the relations quoted for \(MH\) over \(\mathbb Q(v)\) [1310.3878], [2404.05594].

A useful reformulation introduces the idempotent
\[
e=q^{-1}(T_0+1).
\]
Then the algebra is generated by \(e,T_1,\dots,T_{n-1}\) with the ordinary Hecke relations for \(T_i\), together with
\[
e^2=e,\qquad eT_i=T_ie\quad (i\ge2),
\]
and
\[
eT_1eT_1=T_1eT_1e,
\]
\[
eT_1eT_1=T_1eT_1-eT_1e+T_1e+eT_1+e-T_1-1.
\]
This presentation makes the “one extra idempotent” structure completely explicit and is the form most directly comparable with cyclotomic Hecke algebras [1310.3878].

Wan gives a further presentation over an arbitrary commutative ring \(R\) with invertible \(q\), replacing \(T_0\) by a family of idempotents \(P_i\). The first is
\[
P_1=1-q^{-1}(T_0+1),
\]
and recursively
\[
P_i=-P_{i-1}T_{i-1}^{-1}P_{i-1}
=-q^{-1}\bigl(P_{i-1}T_{i-1}P_{i-1}-(q-1)P_{i-1}^2\bigr),
\qquad 2\le i\le n.
\]
These satisfy
\[
P_i^2=P_i,\qquad
P_iP_j=P_jP_i=P_i\quad (1\le j<i\le n),
\]
\[
P_iT_j=T_jP_i\quad (1\le i<j\le n),
\qquad
P_iT_j=T_jP_i=-P_i\quad (1\le j<i\le n).
\]
With these generators Wan constructs a basis indexed by triples \((A,B,w)\), where \(A,B\subset\{1,\dots,n\}\) have equal cardinality and \(w\) belongs to a Young subgroup \(\mathfrak S'_{n-k}\). The basis elements are
\[
T_{(A,B,w)}=T_A\,P_k\,T_w\,T_B^{-1},
\]
and this yields an arbitrary-base-ring presentation of \(\mathscr H_{n,R}(q)\) [2603.00603].

## 3. Cyclotomic realization, center, cocenter, and character theory

A fundamental structural result is that the mirabolic Hecke algebra is a quotient of the Ariki–Koike cyclotomic Hecke algebra \(H_n(1,0)\). In Rosso’s formulation,
\[
R_n\cong H_n(1,0)/I_{(0,2)},
\]
where \(I_{(0,2)}\) is the two-sided ideal generated by the primitive idempotent corresponding to the bipartition \((0,2)\). This quotient description explains why simple modules are indexed not by arbitrary bipartitions, but by those whose second component is a single column [1310.3878].

More precisely, the simple modules of the generic mirabolic Hecke algebra are indexed by pairs \((\lambda,k)\), where \(0\le k\le n\) and \(\lambda\vdash k\); equivalently, by bipartitions of the form
\[
(\lambda,1^{n-k}).
\]
This is the representation-theoretic shadow of the fact that the mirabolic datum consists of a single vector rather than a higher-dimensional auxiliary space [1310.3878].

The algebra admits mirabolic Jucys–Murphy elements. In the \(e\)-presentation they are
\[
L_i=q^{1-i}T_{i-1}\cdots T_1\,e\,T_1\cdots T_{i-1},
\]
and they act diagonally in the seminormal basis attached to standard bitableaux. Rosso proves that the center is exactly the ring of symmetric polynomials in these elements:
\[
Z(R_n)=\mathbb C(q)[L_1,\dots,L_n]^{S_n}.
\]
This is the direct mirabolic analogue of the classical description of the center of the type-\(A\) Hecke algebra [1310.3878].

Wan extends the structural picture from the center to the cocenter. He defines special elements
\[
\widehat T^{(n)}_\mu=P_{n-|\mu|}\,T_{w_{\mu^{\uparrow n}}},
\]
indexed by partitions \(\mu\) of size at most \(n\), and shows that their images form a basis of the cocenter \(H/[H,H]\). Every basis element \(T_{(A,B,w)}\) admits an expansion modulo commutators in terms of these \(\widehat T^{(n)}_\lambda\), with uniquely determined class polynomials \(f_{(A,B,w)}^\lambda(q)\). On this basis Wan defines the character table of \(\mathscr H_n(q)\), proves a Frobenius character formula in the ring of symmetric functions, and derives a recursive Murnaghan–Nakayama rule for the irreducible characters [2603.00603].

The Frobenius formula takes the form
\[
\widetilde q_\mu(x_1,\dots,x_r,1;q)
=
\sum_{(\lambda,k)}
s_\lambda(x_1,\dots,x_r)\,
\chi^{(n)}_{(\lambda,k)}\bigl(\widehat T^{(n)}_\mu\bigr),
\]
with Schur functions on the right and Hall–Littlewood-type symmetric functions on the left. The associated Murnaghan–Nakayama rule removes strips from \(\lambda\) with explicit \(q\)-dependent weights \(\overline{\mathrm{wt}}_{\lambda/\nu}(q)\), giving a recursive computation of the mirabolic character table [2603.00603].

## 4. Dualities: Schur–Weyl, double centralizers, and Howe theory

The geometric role of the mirabolic Hecke algebra is most transparent in mirabolic Schur–Weyl duality. Let
\[
MS_{n,d}=A^G(\mathcal X\times\mathcal X\times V),
\qquad
MH=A^G(\mathcal Y\times\mathcal Y\times V),
\]
and
\[
M_V=A^G(\mathcal X\times\mathcal Y\times V).
\]
Then \(M_V\) is naturally an \((MS_{n,d},MH)\)-bimodule by convolution. For \(n\ge d\), the two actions form a double centralizer pair:
\[
\operatorname{End}_{MS_{n,d}}(M_V)\cong MH,
\qquad
\operatorname{End}_{MH}(M_V)\cong MS_{n,d}.
\]
This is the geometric mirabolic Schur–Weyl duality of type \(A\) [2404.05594].

The same literature relates \(MH\) to the mirabolic quantum group \(MU\). The left \(MU\)-action and right \(MH\)-action on \(M_V\) commute, and for \(n\ge d\) one has
\[
\operatorname{End}_{MU}(M_V)\cong MH,
\qquad
MU\to \operatorname{End}_{MH}(M_V)
\text{ surjective.}
\]
Thus the mirabolic Hecke algebra is the full commutant of the mirabolic quantum-group action on the geometric tensor space [2404.05594].

A complementary description uses the semidirect product
\[
G'=G\ltimes V.
\]
In this model,
\[
MH\cong e_BK(v)[G']e_B,
\]
where \(e_B\) is the Borel idempotent. This realizes the mirabolic Hecke algebra as a corner algebra of the group algebra of \(G\ltimes V\), directly paralleling the familiar type-\(A\) description \(e_BK(v)[G]e_B\) for the ordinary Hecke algebra. The same framework supports mirabolic Howe duality: for \(n\ge m\ge d\), the bimodule
\[
MV_{n|m}\cong MV_{n,d}\otimes_{MH}MV_{m,d}
\]
carries commuting actions of \(MU_n\) and \(MU_m\), and these actions are mutual centralizers [2412.04136].

Wan establishes a distinct, algebraic Schur–Weyl duality with the ordinary quantum group \(U_q(\mathfrak{gl}_r)\). On \(V_{r+1}^{\otimes n}\), the actions of \(U_q(\mathfrak{gl}_r)\) and \(\mathscr H_n(q)\) centralize each other, and one obtains
\[
V_{r+1}^{\otimes n}\cong
\bigoplus_{k=0}^n
\bigoplus_{\substack{\lambda\vdash k\\ \ell(\lambda)\le r}}
L(\lambda)\otimes N^{(n)}_{(\lambda,k)}.
\]
This places the mirabolic Hecke algebra in a see-saw with Jimbo’s classical duality for \(\mathcal H_n(q)\subset \mathscr H_n(q)\) and \(U_q(\mathfrak{gl}_r)\subset U_q(\mathfrak{gl}_{r+1})\) [2603.00603].

## 5. Representation theory

At generic parameter the mirabolic Hecke algebra is semisimple. Rosso’s cyclotomic realization recovers Siegel’s classification of irreducibles, with simple modules indexed by pairs \((\lambda,k)\) or, equivalently, by bipartitions \((\lambda,1^{n-k})\). Restriction and induction admit explicit branching rules, and the mirabolic Jucys–Murphy elements refine these functors in a way analogous to the ordinary Hecke-theoretic picture [1310.3878].

This refinement leads to a \(\mathfrak{gl}_\infty\)-structure on the Grothendieck group of the tower of mirabolic Hecke algebras. Rosso defines exact functors \({}^i\mathrm{Res}\) and \({}^i\mathrm{Ind}\) via generalized eigenspaces of the last Jucys–Murphy element, and from these constructs operators \(e_i,f_i\) on
\[
G(R)=\bigoplus_{n\ge0}K_0(R_n\text{-Mod}).
\]
The resulting module decomposes into countably many copies of the charge-\(0\) level-\(1\) Fock space [1310.3878].

The relation to mirabolic quantum groups sharpens the representation theory further. In the \(n=2\) case, Rosso’s mirabolic quantum \(\mathfrak{sl}_2\) yields a Schur–Weyl dictionary between simple modules of \(MU_v(2)\) and simple modules of the mirabolic Hecke algebra, and this recovers the classification of irreducible \(MH\)-modules. Later work for general \(n\) makes the correspondence explicit: the irreducible \(MU(n)\)-modules \(L_{\lambda,s}\) are paired with irreducible mirabolic Hecke modules \(M^{(\lambda,1^s)}\) through mirabolic quantum Schur–Weyl duality [1509.04790], [2510.07469].

A recurrent feature is that only bipartitions with second component a column occur. This should not be viewed as a notational accident. It reflects the geometry of a single distinguished vector and is precisely the constraint predicted by both the cyclotomic quotient \(H_n(1,0)/I_{(0,2)}\) and the mirabolic Schur–Weyl decomposition [1310.3878].

## 6. Broader context, scope, and open directions

The mirabolic Hecke algebra belongs to a larger family of “mirabolic” constructions in geometric representation theory. Travkin studied convolution and Kazhdan–Lusztig-type bases on \(\mathcal Y\times\mathcal Y\times V\); Finkelberg–Ginzburg introduced categories of mirabolic \(D\)-modules and related them to spherical trigonometric Cherednik algebras; and Bellamy–Ginzburg showed that the spherical trigonometric Cherednik algebra \(U_k\) is morally a mirabolic version of the affine Hecke algebra. This suggests a conceptual continuum between finite mirabolic Hecke algebras, mirabolic character-sheaf phenomena, and Cherednik-theoretic Hamiltonian reduction, but the papers distinguish these objects carefully rather than identifying them outright [1207.1797].

One common misconception is that the mirabolic Hecke algebra is merely the ordinary type-\(A\) Hecke algebra with an ad hoc extra idempotent. The literature does not support that simplification. The defining mixed relations, the orbit parametrization by \((w,\beta)\), the appearance of decorated matrices on partial-flag spaces, and the double-centralizer role of \(MH\) all arise from a genuine enlargement of the underlying geometry from pairs of flags to triples of two flags and a vector [2404.05594].

Another active theme is cellularity. Rosso conjectured that \(R_n(q)\) is a cellular algebra, proposed a strategy via the cellular basis of the cyclotomic Hecke algebra \(H_n(1,0;q)\), and verified the conjecture for \(R_2(q)\). A full proof for all \(n\) was not available in that work [1310.3878]. Wan’s arbitrary-ring presentation and cocenter basis suggest that structural questions of this sort remain central to the subject [2603.00603].

The recent direction of the field is therefore twofold. One direction strengthens the internal structure of \(MH\): new presentations, cocenters, character formulas, and recursive character theory. The other embeds \(MH\) into larger duality frameworks: geometric mirabolic Schur–Weyl duality, mirabolic Howe duality, and algebraic duality with \(U_q(\mathfrak{gl}_r)\). Together these developments establish the mirabolic Hecke algebra as a stable object of type-\(A\) representation theory, with its own orbit combinatorics, central and cocentral structure, and a network of dualities paralleling—but not duplicating—the classical theory [2412.04136], [2603.00603].

Source: https://www.emergentmind.com/topics/mirabolic-hecke-algebra