---
title: Hermite Reciprocity in SL₂ Representations
url: https://www.emergentmind.com/topics/hermite-reciprocity
type: topic
---

# Hermite Reciprocity in SL₂ Representations

Searching arXiv for recent papers on Hermite reciprocity and related work.
Hermite reciprocity is a phenomenon unique to representations of $\mathrm{SL}_2$ that relates plethysms of binary forms, most classically through the isomorphism
$\mathrm{Sym}^a(\mathrm{Sym}^b U)\cong \mathrm{Sym}^b(\mathrm{Sym}^a U)$ for $U\cong \mathbf{C}^2$. In characteristic $0$ it appears as an isomorphism of $\mathrm{SL}_2$-modules; over an arbitrary base field it admits natural formulations involving exterior and divided powers; and in recent work it has been lifted from representation theory to isomorphisms of complexes, where it governs self-duality phenomena in syzygy theory, secant geometry, generalized Eagon–Northcott complexes, and the ideal-theoretic structure of varieties of powers of binary forms [2106.04495] [2504.07184] [2602.15175].

## 1. Classical statement and representation-theoretic setting

Let $U=\mathbf{C}^2$ with the standard $\mathrm{SL}(U)=\mathrm{SL}_2$ action. For $b\ge 0$, $\mathrm{Sym}^b U$ is the space of binary forms of degree $b$, and plethysm refers to modules such as $\mathrm{Sym}^a(\mathrm{Sym}^b U)$. The classical theorem, proved by Hermite in 1854, states that for all $a,b\ge 0$,
\[
\mathrm{Sym}^a(\mathrm{Sym}^b U)\cong \mathrm{Sym}^b(\mathrm{Sym}^a U)
\]
as $\mathrm{SL}_2$-modules, equivalently as $\mathrm{GL}_2$-modules with central character. This symmetry is specific to rank $2$; for higher-rank groups, plethysm need not commute [2602.15175].

A more functorial formulation uses exterior and divided powers. For integers $m,n$ with $n\ge m\ge 0$, there are canonical $\mathrm{SL}_2$-equivariant isomorphisms
\[
\Lambda^m(\mathrm{Sym}^n U)\cong \mathrm{Sym}^{n-m+1}(D^m U), \qquad
D^m(\mathrm{Sym}^{n-m} U)\cong \mathrm{Sym}^{n-m}(D^m U).
\]
These statements hold over an arbitrary base field by working with divided powers $D^m$, and in characteristic $0$ one may replace $D^m$ by $\mathrm{Sym}^m$ because complete reducibility holds. All constructions are $\mathrm{SL}_2$-equivariant, and $\mathrm{GL}$-equivariance can be restored by suitable determinant twists [2106.04495].

Two points are often conflated. First, in the $\mathrm{SL}_2$ setting the reciprocity is not merely an equality of formal characters: it is an isomorphism of modules. Second, the phenomenon is not a generic plethystic symmetry; it is a rank-$2$ exception, and much of the modern literature is devoted to identifying the geometric and homological structures that explain why this exception persists in increasingly derived forms [2106.04495].

## 2. Character formulas, combinatorics, and explicit decompositions

For $\mathrm{SL}_2$, one effective way to encode characters uses a single formal variable $q$. After specializing the diagonal torus weights $(x_1,x_2)$ to $(q,1)$, one has
\[
\mathrm{ch}(\mathrm{Sym}^n U)=1+q+\cdots+q^n=:[n+1]_q,
\]
together with the formulas
\[
\mathrm{ch}(\Lambda^m(\mathrm{Sym}^n U))=\binom{n+1}{m}_q,\qquad
\mathrm{ch}(\mathrm{Sym}^m(\mathrm{Sym}^n U))=\binom{m+n}{m}_q.
\]
These identities make classical character-theoretic proofs efficient: one computes both sides of Hermite reciprocity and then invokes Schur’s lemma [2602.15175].

A complementary combinatorial description decomposes
\[
\mathrm{Sym}^m(\mathrm{Sym}^n V)\cong
\bigoplus_{k=0}^{\lfloor mn/2\rfloor}\mathrm{Sym}^{mn-2k}(V)^{\oplus \mu_{m,n}(k)},
\]
where $\mu_{m,n}(k)$ is the number of partitions of $k$ whose Ferrers diagram fits inside an $m\times n$ rectangle. This description is symmetric in $(m,n)$ by construction, and parity forces $\mathrm{Sym}^r(V)$ to occur only for $r\equiv mn \bmod 2$ [2106.04495].

| Parameters | Decomposition | Reciprocity partner |
|---|---|---|
| $(2,3)$ | $\mathrm{Sym}^2(\mathrm{Sym}^3 U)\cong \mathrm{Sym}^6 U\oplus \mathrm{Sym}^2 U$ | $\mathrm{Sym}^3(\mathrm{Sym}^2 U)\cong \mathrm{Sym}^6 U\oplus \mathrm{Sym}^2 U$ |
| $(2,4)$ | $\mathrm{Sym}^2(\mathrm{Sym}^4 U)\cong \mathrm{Sym}^8 U\oplus \mathrm{Sym}^4 U\oplus \mathrm{Sym}^0 U$ | $\mathrm{Sym}^4(\mathrm{Sym}^2 U)\cong \mathrm{Sym}^8 U\oplus \mathrm{Sym}^4 U\oplus \mathrm{Sym}^0 U$ |
| $(3,3)$ | $\mathrm{Sym}^3(\mathrm{Sym}^3 V)\cong \mathrm{Sym}^9 V\oplus \mathrm{Sym}^5 V\oplus \mathrm{Sym}^1 V$ | self-symmetric |

The classical literature also supplies an explicit invariant-theoretic route through transvectants. If $F\in \mathrm{Sym}^d U$ and $G\in \mathrm{Sym}^e U$ are binary forms, then the $r$-th transvectant $(F,G)_r\in \mathrm{Sym}^{d+e-2r}U$ is defined by polarization and contraction with the $\mathrm{SL}_2$-invariant bilinear form on $U$. In particular, the first transvectant agrees, up to scalar, with the Jacobian determinant of the gradients. Hermite’s construction uses suitable transvectant operators to produce an explicit intertwiner between the plethysms [2602.15175].

## 3. Cohomological realizations and Schwarzenberger bundles

A geometric realization begins with the Hilbert scheme $\mathrm{Hilb}^m(\mathbf{P}^1)\cong \mathbf{P}^m$ and the incidence correspondence
\[
Z=\{([f],[p])\in \mathbf{P}^m\times \mathbf{P}^1\mid f(p)=0\}.
\]
With projections $\pi_1,\pi_2$, one defines the Schwarzenberger bundle $E_m=\pi_{1*}(\mathcal{O}_Z(0,n))$ on $\mathbf{P}^m$. It fits into the exact sequence
\[
0\to \mathrm{Sym}^n U\otimes \mathcal{O}_{\mathbf{P}^m}(-1)\to
\mathrm{Sym}^{n+m} U\otimes \mathcal{O}_{\mathbf{P}^m}\to E_m\to 0,
\]
with $\mathrm{rank}(E_m)=m$ and $\det(E_m)=\mathcal{O}_{\mathbf{P}^m}(n+1)$. Taking the $m$-th exterior power and computing cohomology yields Hermite reciprocity, and the algebraic and geometric constructions agree compatibly across all $n\ge m$ [2106.04495].

The exterior powers of Schwarzenberger bundles have supernatural cohomology. For $0<i<m$, one has
\[
\Lambda^i E_m \cong p_*\mathcal{O}_{\mathbf{P}^i\times \mathbf{P}^{m-i}}(d+m-i+1,0),
\]
where $p:\mathbf{P}^i\times \mathbf{P}^{m-i}\to \mathbf{P}^m$ is multiplication of binary forms. The associated root sequence is
\[
-1,-2,\ldots,-(m-i),-(m-i+d+2),\ldots,-(m+d+1).
\]
This identifies $\Lambda^iE_m$ as a special case of the Eisenbud–Schreyer construction [2106.04495].

The same geometry controls secant varieties of rational normal curves. If $C_n\subset \mathbf{P}^n$ is the degree-$n$ Veronese image of $\mathbf{P}^1$, then the $k$-secant variety $\Sigma$ is resolved by the total space of the Schwarzenberger bundle $E_k$ on $\mathbf{P}^k$. The coordinate ring is resolved by the Eagon–Northcott complex of a Hankel matrix, and the resulting secant variety is normal, arithmetically Cohen–Macaulay, and has rational singularities. At the module-theoretic level, Hermite reciprocity becomes equivalent to a canonical self-duality
\[
M_{-1}\cong M_{n-2k+1}
\]
between rank-one maximal Cohen–Macaulay modules on the affine cone of the secant variety; on degree-$0$ generators this restricts to
\[
\Lambda^k(\mathrm{Sym}^{n-k}U)\cong \mathrm{Sym}^{n-2k+1}(D^kU).
\]
Thus the reciprocity isomorphism is realized as the generator of a $1$-dimensional space of sections
\[
H^0(\mathbf{P}^k,\mathrm{Sym}^k(E)\otimes \mathcal{O}_{\mathbf{P}^k}(-n+2k-2)).
\]
This furnishes a geometric explanation for the relation between plethysm and self-duality [2106.04495].

## 4. Derived Hermite reciprocity

Recent work lifts the classical theorem from modules to complexes. Fix $a,b\ge 1$ and set $d=ab$. On $\mathbf{P}(\mathrm{Sym}^b U)$ there is an $\mathrm{SL}_2$-equivariant sheaf map
\[
\phi:\mathrm{Sym}^{d+b-2}U\otimes \mathcal{O}_{\mathbf{P}^b}(-1)\to \mathrm{Sym}^dU\otimes \mathcal{O}_{\mathbf{P}^b},
\]
whose cokernel is $\mathcal{O}_{\mathbf{P}^b}(a)$ and whose kernel is a rank $b-1$ bundle $F$. This $\phi$ is the universal syzygy among the sections defining the embedding
\[
\nu_a:\mathbf{P}^b\to \mathbf{P}^d,\qquad [G]\mapsto [G^a].
\]
From $\phi$ one constructs symmetric power complexes $S_r^\bullet=\mathrm{Sym}^r(\phi)$ and exterior power complexes $W_r^\bullet=\Lambda^r(\phi^\vee(-1))$, with cohomology sheaves described by exterior powers of $F$ and $F^\vee$ [2602.15175].

The derived form of Hermite reciprocity states that for all $a,b\ge 1$,
\[
S_{b-1}^\bullet \cong W_{b-1}^\bullet
\]
as complexes on $\mathbf{P}^b$, equivariantly for $\mathrm{GL}(U)$. Termwise, this yields
\[
\Lambda^i(\mathrm{Sym}^{d+b-2}U)\otimes \mathrm{Sym}^{b-1-i}(\mathrm{Sym}^dU)
\cong
\mathrm{Sym}^i(\mathrm{Sym}^dU)\otimes \Lambda^{b-1-i}(\mathrm{Sym}^{d+b-2}U),
\]
and the compatibility of differentials is highly nontrivial. Character computations justify the termwise isomorphism by products of $q$-binomial coefficients, but the proof of the complex isomorphism uses a geometric realization as derived pushforwards of Koszul complexes along the graph of $\nu_a$, together with Beilinson’s equivalence and a self-duality coming from the Cartier divisor
\[
\mathcal{O}_\Gamma(0,b-1)\cong \mathcal{O}_\Gamma(d,-1).
\]
In this sense, the derived theorem is not merely a restatement of classical plethysm symmetry; it is a self-duality statement internal to a syzygetic construction [2602.15175].

## 5. Foulkes–Howe maps, powers of binary forms, and determinantal equations

The embedding $\nu_a$ induces generalized Foulkes–Howe maps
\[
\alpha_k:\mathrm{Sym}^k(\mathrm{Sym}^{ab}U)\to \mathrm{Sym}^{ak}(\mathrm{Sym}^bU),\qquad k\ge 0.
\]
A principal consequence of derived Hermite reciprocity is maximal rank: $\alpha_k$ is injective for $k\le b$ and surjective for $k\ge b$; in particular,
\[
\alpha_b:\mathrm{Sym}^b(\mathrm{Sym}^dU)\to \mathrm{Sym}^d(\mathrm{Sym}^bU)
\]
is an isomorphism. A common misconception is that this distinguished geometric map must coincide with the classical Hermite isomorphism. It does not in general: for $(a,b)=(2,4)$, the paper compares images of highest weight vectors and finds
\[
p(\alpha_2)=(1,-1/3,4/3),\qquad p(h_{2,4})=(1,1/4,1/2).
\]
The abstract reciprocity and the geometric Foulkes–Howe map therefore encode different structures [2602.15175].

The same machinery resolves the variety
\[
X=X_{(a^b)}=\{[F]\in \mathbf{P}(\mathrm{Sym}^dU)\mid F=G^a \text{ for some } G\in \mathrm{Sym}^bU\},
\]
the locus of $a$-th powers of binary $b$-ics. Define the bilinear Jacobian/transvectant map
\[
\omega(F,G)=
\det\begin{bmatrix}
\partial F/\partial x_1 & \partial F/\partial x_2\\
\partial G/\partial x_1 & \partial G/\partial x_2
\end{bmatrix}.
\]
Then $\omega(F,G)=0$ if and only if $[F]=[G^a]$, so the maximal minors of the sheaf map
\[
\omega:\mathrm{Sym}^bU\otimes \mathcal{O}_{\mathbf{P}^d}(-1)\to \mathrm{Sym}^{d+b-2}U\otimes \mathcal{O}_{\mathbf{P}^d}
\]
cut out $X$ set-theoretically. More strongly, for $a,b\ge 2$, the homogeneous ideal $I(X)$ is generated in degree $b+1$ by those maximal minors; its minimal free resolution is linear and has projective dimension $d-1$; and the number of minimal generators is
\[
\dim I(X)_{b+1}=\binom{d+b+1}{b+1}-\binom{d+a+b}{b}.
\]
The determinantal presentation is encoded by the exact sequence
\[
\Lambda^{b+1}(\mathrm{Sym}^{d+b-2}U)\to \mathrm{Sym}^{b+1}(\mathrm{Sym}^dU)\to \mathrm{Sym}^{a(b+1)}(\mathrm{Sym}^bU)\to 0,
\]
whose middle map is $\alpha_{b+1}$ and whose left map is the universal minors map $\beta$; here $\ker(\alpha_{b+1})=\mathrm{Im}(\beta)$ [2602.15175].

Examples align with classical invariant theory. For $b=2$, $X$ is cut out by cubics; for $b=3$, by quartics; and for $b=4$, by quintics, where the maximal minors of $\omega$ agree up to scalar with Hilbert’s classical degree-$5$ transvectant equations. This places Hermite reciprocity inside a concrete determinantal and syzygetic framework rather than only a character-theoretic one [2602.15175].

## 6. New reciprocity maps, generalized Eagon–Northcott complexes, and current directions

A further development concerns generalized Eagon–Northcott complexes associated to Koszul-type maps. Given a map $\phi:F\to G$ of free modules, one has linear complexes $\mathrm{Sym}^i(\phi)$ and $\Lambda^i(\phi)$, and a characteristic-free duality
\[
\mathrm{Sym}^i(\phi)\cong (\Lambda^i(\phi^*))^*.
\]
Splicing these pieces gives generalized Eagon–Northcott complexes $C_i(\phi)$ with duality
\[
C_i(\phi)\cong C_{f-g-i}(\phi)^*.
\]
In the generic determinantal case these complexes are not self-dual; self-duality is therefore a special phenomenon requiring specialization [2504.07184].

The 2025 work identifies such a specialization in the Koszul setting. Let $V_0$ be a complex vector space of dimension $b+1$, let $V_1\subset \Lambda^2V_0$ have dimension $2b-1$, and assume that the restricted sequence
\[
V_1\otimes S \to V_0\otimes S \to S\to 0
\]
has middle homology of finite length. Then $\mathrm{Sym}^{b-1}(\phi|_{V_1\otimes S})$ is self-dual. In the $\mathrm{SL}_2$ specialization $V_0=\mathrm{Sym}^bU$ and $V_1=\mathrm{Sym}^{2b-2}U\subset \Lambda^2(\mathrm{Sym}^bU)$, this produces a new Hermite reciprocity isomorphism
\[
\psi_{b,0}(V_1):\Lambda^b(\mathrm{Sym}^{2b-2}U)\to \mathrm{Sym}^{b-1}(\mathrm{Sym}^bU)\otimes \det(\mathrm{Sym}^bU),
\]
described as not previously defined in the literature. For $b=3$, the paper gives explicit matrices for this map and for an older Hermite isomorphism and shows that they differ on three $2\times 2$ diagonal blocks [2504.07184].

This new map is used to prove self-duality by reducing, via the BGG correspondence, to a single lowest-degree check. In the same $\mathrm{SL}_2$ specialization, the middle homology is the Weyman module used in proofs of the generic Green’s conjecture. The broader significance is that Hermite reciprocity now appears in three interconnected guises: as classical plethysm symmetry, as self-duality of modules on secant cones, and as self-duality of complexes controlling Koszul and Weyman modules. This suggests broader families of complex isomorphisms and further higher-rank analogues, although the strongest reciprocity phenomenon remains special to rank $2$. Current open directions include extending the self-duality to other ranks, determining minimal free resolutions of the ideals $I_{a,b}$ for all $b$, and proving the conjectural regularity formula for powers $I_{a,b}^j$ [2504.07184] [2602.15175].

Source: https://www.emergentmind.com/topics/hermite-reciprocity