Field-independent Kronecker-plethysm isomorphisms

Christian Ikenmeyer University of Warwick [email protected], Heidi Omar University of Warwick [email protected], Dimitrios Tsintsilidas University of Warwick [email protected] · 2025

Read Field-independent Kronecker-plethysm isomorphisms by Christian Ikenmeyer University of Warwick [email protected], Heidi Omar University of Warwick [email protected], Dimitrios Tsintsilidas University of Warwick [email protected] (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2509-10069.txt.

Introduction

Christian Ikenmeyer and 2 co-authors build a bridge between two symmetry spaces in a 2025 paper. The bridge works over any field. It links a space of invariant tensors to a plethysm space. It respects the action of the special linear group in two dimensions. Such a map was known only in characteristic zero. Even there it was known only through character counts.

Their map is written down in full. It needs no assumption on the field at all. It joins a small web of such maps. Hermite reciprocity and Hodge duality sit in that web. The Wronskian map completes the set. The authors add a short proof about one older map. Over the standard basis, that map is triangular. Its diagonal holds only ones. Matching coefficient patterns sit behind the new bridge.

Catalog record

CREATORChristian Ikenmeyer University of Warwick [email protected], Heidi Omar University of Warwick [email protected], Dimitrios Tsintsilidas University of Warwick [email protected]
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE53 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2509-10069

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Field-independent Kronecker-plethysm isomorphisms Christian Ikenmeyer University of Warwick [email protected], Heidi Omar University of Warwick [email protected], Dimitrios Tsintsilidas University of Warwick [email protected] Abstract We construct an explicit field-independent SL 2 -equivariant isomorphism between an invariant space of tensors and a plethysm space. The existence of such an isomorphism was only known in characteristic 0, and only indirectly via character theory. Our isomorphism naturally extends the web of field-independent isomorphisms given by Hermite reciprocity, Hodge duality, and the Wronskian isomorphism. This is a characteristic free generalization of a classical situation in characteristic zero: certain rectangular Kronecker coefficients coincide with certain plethysm coefficients, and their non-negativiy proves the unimodality of the q q -binomial coefficient. We also give a short combinatorial field-independent proof that the Hermite reciprocity map over the standard basis is a triangular matrix with 1s on the main diagonal. † † footnotetext: For the purpose of open access, the authors have applied a Creative Commons Attribution (CC-BY) licence to any Author Accepted Manuscript version arising from this submission. 1 Introduction The classical Hermite reciprocity law states that \textup{Sym}^{m}(\textup{Sym}^{\ell}\mathbb{C}^{2}) and \textup{Sym}^{\ell}(\textup{Sym}^{m}\mathbb{C}^{2}) are isomorphic \textsf{{GL}}_{2}(\mathbb{C}) -representations, where \textup{Sym}^{m}W denotes the symmetric power. This statement is false over arbitrary fields, as was shown in [ Kou90 ] . In fact, proper duals have to be taken, see ( 1 ) below. For any field \mathbb{F} , let W be a \textsf{{GL}}_{2}(\mathbb{F}) -representation, and let \textup{Sym}^{m}W=\big(({\textstyle\bigotimes}^{\bullet}W)/\langle x\otimes y-y\otimes x\rangle\big)_{m} denote its m -th symmetric power. Let \mathfrak{S}_{m} denote the symmetric group on m symbols. Let \textup{Sym}_{m}W=({\textstyle\bigotimes}^{m}W)^{\mathfrak{S}_{m}} denote the m -th divided power of W , i.e., the vector space of \mathfrak{S}_{m} -invariant order m tensors. The representations \textup{Sym}^{m}W and \textup{Sym}_{m}W are dual to each other (see Lemma 1.1 below). The self-duality of \textsf{{SL}}_{2}(\mathbb{C}) -representations makes \textup{Sym}^{m}W and \textup{Sym}_{m}W isomorphic \textsf{{SL}}_{2}(\mathbb{C}) -representations over \mathbb{C} , which makes it difficult to see the field-independent structure. One gets a Hermite reciprocity isomorphism \textsf{{R}}_{m,\ell} in a field-independent way after choosing the duals corrrectly: \textsf{{R}}_{m,\ell}:\textup{Sym}_{m}\textup{Sym}^{\ell}\mathbb{F}^{2}\ \stackrel{{\scriptstyle\sim}}{{\longrightarrow}}\ \textup{Sym}^{\ell}\textup{Sym}_{m}\mathbb{F}^{2}, (1) see [ AFP+19 ] and [ MW22 ] , later again discussed in [ RS21 ] . The isomorphism \textsf{{R}}_{m,\ell} is defined via field-independent versions of classical isomorphisms, as depicted in Figure 1 : \textsf{{R}}_{m,\ell}=\textsf{{W}}^{\star}_{\ell,m}\circ\textsf{{D}}_{m,\ell}\circ\textsf{{W}}_{m,\ell} . In this paper, we add new explicit isomorphisms to this picture that connect plethysm spaces (i.e., compositions of \textup{Sym}^{\bullet} or \textup{Sym}_{\bullet} , such as \textup{Sym}_{m}\textup{Sym}^{\ell}\mathbb{F}^{2} ) with invariant spaces of tensors as follows. The isomorphisms: \textup{Sym}_{m}\textup{Sym}^{\ell}\mathbb{F}^{2} {\textstyle\bigwedge}^{m}\textup{Sym}^{\ell+m-1}\mathbb{F}^{2} \textup{Sym}^{\ell}\textup{Sym}_{m}\mathbb{F}^{2} {\textstyle\bigwedge}^{\ell}\textup{Sym}_{\ell+m-1}\mathbb{F}^{2} \textsf{{W}}_{m,\ell} \textsf{{D}}_{m,\ell}=\textsf{{D}}_{\ell,m}^{\star} \textsf{{W}}_{\ell,m}^{\star} \textsf{{R}}_{m,\ell}=\textsf{{R}}^{\star}_{\ell,m} \big(\textup{Sym}_{\ell m}(\mathbb{F}^{\ell\times\ell\times 2})\big)^{\textsf{{SL}}_{\ell}({\overline{\mathbb{F}}})\times\textsf{{SL}}_{\ell}({\overline{\mathbb{F}}})} \big(\textup{Sym}^{\ell m}(\mathbb{F}^{m\times m\times 2})\big)^{\textsf{{SL}}_{m}({\overline{\mathbb{F}}})\times\textsf{{SL}}_{m}({\overline{\mathbb{F}}})} \textsf{{K}}_{\ell,m} \textsf{{K}}^{\star}_{m,\ell} \textsf{{I}}_{m,\ell}=\textsf{{I}}_{\ell,m}^{\star} The dual situation. The same maps appear, just with swapped parameters: \textup{Sym}^{m}\textup{Sym}_{\ell}\mathbb{F}^{2} {\textstyle\bigwedge}^{m}\textup{Sym}_{\ell+m-1}\mathbb{F}^{2} \textup{Sym}_{\ell}\textup{Sym}^{m}\mathbb{F}^{2} {\textstyle\bigwedge}^{\ell}\textup{Sym}^{\ell+m-1}\mathbb{F}^{2} \textsf{{W}}^{\star}_{m,\ell} \textsf{{D}}_{\ell,m}=\textsf{{D}}_{m,\ell}^{\star} \textsf{{W}}_{\ell,m} \textsf{{R}}_{\ell,m}=\textsf{{R}}^{\star}_{m,\ell} \big(\textup{Sym}^{\ell m}(\mathbb{F}^{\ell\times\ell\times 2})\big)^{\textsf{{SL}}_{\ell}({\overline{\mathbb{F}}})\times\textsf{{SL}}_{\ell}({\overline{\mathbb{F}}})} \big(\textup{Sym}_{\ell m}(\mathbb{F}^{m\times m\times 2})\big)^{\textsf{{SL}}_{m}({\overline{\mathbb{F}}})\times\textsf{{SL}}_{m}({\overline{\mathbb{F}}})} \textsf{{K}}_{\ell,m}^{\star} \textsf{{K}}_{m,\ell} \textsf{{I}}_{\ell,m}=\textsf{{I}}_{m,\ell}^{\star} Figure 1: Commutative diagrams of field-independent equivariant isomorphisms and their duals. The map \textsf{{R}}_{m,\ell} is Hermite reciprocity, \textsf{{W}}_{m,\ell} is the Wronskian isomorphism, \textsf{{D}}_{m,\ell} is the Hodge duality, \textsf{{K}}_{m,\ell} is our Kronecker-plethysm isomorphism, and \textsf{{I}}_{m,\ell} is our isomorphism between tensor invariant spaces. Let {\overline{\mathbb{F}}} denote the algebraic closure of \mathbb{F} , and fix an embedding \mathbb{F}\subseteq{\overline{\mathbb{F}}} . We have \mathbb{F}^{n}\subseteq{\overline{\mathbb{F}}}^{n} , \textup{Sym}^{m}\mathbb{F}^{n}\subseteq\textup{Sym}^{m}{\overline{\mathbb{F}}}^{n} , etc. The group G_{\ell}:=\textsf{{GL}}_{\ell}({\overline{\mathbb{F}}})\times\textsf{{GL}}_{\ell}({\overline{\mathbb{F}}})\times\textsf{{GL}}_{2}({\overline{\mathbb{F}}}) acts on the tensor product {\overline{\mathbb{F}}}^{\ell\times\ell\times 2}:={\overline{\mathbb{F}}}^{\ell}\otimes{\overline{\mathbb{F}}}^{\ell}\otimes{\overline{\mathbb{F}}}^{2} via (g_{1},g_{2},g_{3})(v_{1}\otimes v_{2}\otimes v_{3})=g_{1}(v_{1})\otimes g_{2}(v_{2})\otimes g_{3}(v_{3}) and extended {\overline{\mathbb{F}}} -linearly. This action lifts to the tensor algebra {\textstyle\bigotimes}^{\bullet}({\overline{\mathbb{F}}}^{\ell\times\ell\times 2}) via g(w_{1}\otimes w_{2}\otimes\cdots\otimes w_{d}):=(gw_{1})\otimes\cdots\otimes(gw_{d}) and extended {\overline{\mathbb{F}}} -linearly, for g\in G_{\ell} . This action induces a linear action of G_{\ell} on \textup{Sym}^{\bullet}({\overline{\mathbb{F}}}^{\ell\times\ell\times 2}) and on every \textup{Sym}_{m}({\overline{\mathbb{F}}}^{\ell\times\ell\times 2}) . For any G_{\ell} -representation W , let W^{\textsf{{SL}}_{\ell}({\overline{\mathbb{F}}})\times\textsf{{SL}}_{\ell}({\overline{\mathbb{F}}})} denote the space of invariants under the action of the group \textsf{{SL}}_{\ell}({\overline{\mathbb{F}}})\times\textsf{{SL}}_{\ell}({\overline{\mathbb{F}}}) , interpreted as a subgroup of G_{\ell} via the embedding (g_{1},g_{2})\mapsto(g_{1},g_{2},\textup{id}_{{\overline{\mathbb{F}}}^{2}}) . Define the \mathbb{F} -vector space (\textup{Sym}^{\bullet}(\mathbb{F}^{\ell\times\ell\times 2}))^{\textsf{{SL}}_{\ell}({\overline{\mathbb{F}}})\times\textsf{{SL}}_{\ell}({\overline{\mathbb{F}}})}\,\ :=\ \,(\textup{Sym}^{\bullet}({\overline{\mathbb{F}}}^{\ell\times\ell\times 2}))^{\textsf{{SL}}_{\ell}({\overline{\mathbb{F}}})\times\textsf{{SL}}_{\ell}({\overline{\mathbb{F}}})}\ \cap\ \textup{Sym}^{\bullet}(\mathbb{F}^{\ell\times\ell\times 2}). (2) We establish the explicitly Kronecker-plethysm \textsf{{GL}}_{2}(\mathbb{F}) -isomorphism \textsf{{K}}_{m,\ell}:\big(\textup{Sym}^{\ell m}(\mathbb{F}^{\ell\times\ell\times 2})\big)^{\textsf{{SL}}_{\ell}({\overline{\mathbb{F}}})\times\textsf{{SL}}_{\ell}({\overline{\mathbb{F}}})}\ \stackrel{{\scriptstyle\sim}}{{\longrightarrow}}\ \textup{Sym}^{m}\textup{Sym}_{\ell}\mathbb{F}^{2}. Previously, the existence of such an isomorphism was only known in characteristic zero, and only via the character theory of the symmetric group, see [ PP13 , PP14 ] . We show in § 6 via a direct calculation that the algebraic closure in the definition is necessary, i.e., we give examples where \big(\textup{Sym}^{\ell m}(\mathbb{F}^{\ell\times\ell\times 2})\big)^{\textsf{{SL}}_{\ell}(\mathbb{F})\times\textsf{{SL}}_{\ell}(\mathbb{F})}\ \not\sim\ \textup{Sym}^{m}\textup{Sym}_{\ell}\mathbb{F}^{2} . From Figure 1 , we can now obtain an explicit \textsf{{GL}}_{2}(\mathbb{F}) -isomorphism between invariant spaces of tensors \textsf{{I}}_{m,\ell}:\big(\textup{Sym}_{\ell m}(\mathbb{F}^{\ell\times\ell\times 2})\big)^{\textsf{{SL}}_{\ell}({\overline{\mathbb{F}}})\times\textsf{{SL}}_{\ell}({\overline{\mathbb{F}}})}\to\big(\textup{Sym}^{\ell m}(\mathbb{F}^{m\times m\times 2})\big)^{\textsf{{SL}}_{m}({\overline{\mathbb{F}}})\times\textsf{{SL}}_{m}({\overline{\mathbb{F}}})} via \textsf{{I}}_{m,\ell}:=(\textsf{{K}}_{\ell,m})^{-1}\circ\textsf{{R}}_{m,\ell}\circ(\textsf{{K}}_{m,\ell}^{\star})^{-1} . Again, the existence of such an isomorphism was only known in characteristic zero, and only via the character theory of the symmetric group and Schur-Weyl duality, see for example [ Ike12 , Lem. 4.4.7] . 1.1 Lemma . For a group G and every G -module V over a field \mathbb{F} , we have an isomorphism of G -modules (\textup{Sym}^{n}V)^{\star}\to\textup{Sym}_{n}V^{\star} and ({\textstyle\bigwedge}^{n}V)^{\star}\cong{\textstyle\bigwedge}^{n}V^{\star} . Proof. The first isomorphism is proved in [ McD21 , Prop. 3.7] . The second isomorphism is proved in [ MW22 , Lem. 3.1] . ∎ 1.1 Representation theoretic decompositions and q -binomial coefficients A partition \lambda=(\lambda_{1},\ldots,\lambda_{k}) is a finite list of nonincreasing strictly positive natural numbers. The number l(\lambda)=k is called the length of \lambda . For k>l(\lambda) , we write \lambda_{k}=0 . We write |\lambda|=\sum_{i}\lambda_{i} . We write \lambda\vdash_{n}N if \lambda is a partition with |\lambda|=N and l(\lambda)\leq n . We write \lambda\vdash_{n} if l(\lambda)\leq n with no restriction on |\lambda| . We write (\ell^{m})=(\ell,\ell,\ldots,\ell)\vdash\ell m . We write \lambda\subseteq\mu if \forall i:\lambda_{i}\leq\mu_{i} . To every partition \lambda we associate its Young diagram, which is a top-left justified set of boxes, \lambda_{i} boxes in the i -th row. For example, the Young diagram to \lambda=(4,2,2,1) is . We will always use gray boxes for Young diagrams in this paper, which makes it easier to draw the surrounding grid of potential boxes. Let \mathscr{P}_{k}(\ell,m):=\{\lambda\vdash k,\ \lambda\subseteq(m^{\ell})\} , and let p_{k}(\ell,m):=|\mathscr{P}_{k}(\ell,m)| . Note that p_{k}(\ell,m)=p_{k}(m,\ell) , which can be seen via transposing the partitions, i.e., reflecting the Young diagram at the main diagonal. Let \lambda^{T} denote the partition to the transposed Young diagram of \lambda , for example (4,2,2,1)^{T}=(4,3,1,1) . The irreducible polynomial representations S^{\lambda}\mathbb{C}^{n} of the general linear group \textsf{{GL}}_{n}(\mathbb{C}) are indexed by partitions \lambda\vdash_{n} , see for example [ FH13 ] . The plethysm coefficient a_{\nu}(m[\ell]) is the multiplicity of S^{\nu}\mathbb{C}^{n} in \textup{Sym}^{m}\textup{Sym}^{\ell}\mathbb{C}^{n} , which is independent of n , provided l(\nu)\leq n . Finding a nonnegative combinatorial interpretation for the plethysm coefficient is problem 9 in [ Sta00 ] . For three partitions \lambda\vdash_{k}d , \mu\vdash_{\ell}d , \nu\vdash_{n}d , the Kronecker coefficient \textsf{{k}}(\lambda,\mu,\nu) is the multipliciy of S^{\lambda}\mathbb{C}^{k}\otimes S^{\mu}\mathbb{C}^{\ell}\otimes S^{\nu}\mathbb{C}^{n} in the \textsf{{GL}}_{k}(\mathbb{C})\times\textsf{{GL}}_{\ell}(\mathbb{C})\times\textsf{{GL}}_{n}(\mathbb{C}) -representation \textup{Sym}^{d}(\mathbb{C}^{k}\otimes\mathbb{C}^{\ell}\otimes\mathbb{C}^{n}) . Finding a nonnegative combinatorial interpretation for the Kronecker coefficient is problem 10 in [ Sta00 ] . The special case where k=\ell , d=m\ell for some m , \lambda=\mu=(m^{\ell}) , is called the rectangular Kronecker coefficient \textsf{{k}}((m^{\ell}),(m^{\ell}),\nu) . It equals the multiplicity of S^{\nu}\mathbb{C}^{n} in the \textsf{{GL}}_{n}(\mathbb{C}) -representation \textup{Sym}^{d}(\mathbb{C}^{\ell}\otimes\mathbb{C}^{\ell}\otimes\mathbb{C}^{n})^{\textsf{{SL}}_{\ell}\times\textsf{{SL}}_{\ell}} , provided l(\nu)\leq n . Although plethysm and Kronecker coefficients seem unrelated at first, several equalities, inequalities, and common constructions are known [ Man11 , Ike12 , IP17 , IMW17 , FI20 ] . Both coefficients play an important role in geometric complexity theory, see for example [ BLMW11 , Bür16 , Bür24 ] . In this paper, we focus on the case n=2 . In this case, both coefficients coincide: \textsf{{k}}((m^{\ell}),(m^{\ell}),(\ell m-k,k))=a_{(\ell m-k,k)}(\ell[m])=b_{k}(\ell,m), (3) where b_{k}(\ell,m)=p_{k}(\ell,m)-p_{k-1}(\ell,m) .

Read the complete text