Thrall’s Higher Lie Characters: Decompositions & Applications

• Thrall’s higher Lie characters are the symmetric-group, Frobenius, and GL(V)-module characters arising from the Poincaré–Birkhoff–Witt decomposition of tensor algebras.
• The approach employs induced characters, cyclic sieving, and tableau combinatorics to obtain explicit Schur expansions and determine multiplicities.
• Applications range from combinatorial interpretations of descent statistics to topological realizations, opening new directions in algebraic representation theory.

Searching arXiv for papers on Thrall’s higher Lie characters and related results. Thrall’s higher Lie characters are the symmetric-group characters, Frobenius characteristics, and $GL(V)$-module characters attached to the Poincaré–Birkhoff–Witt decomposition of the tensor algebra through the free Lie algebra. For $V=\mathbb C^N$, the tensor algebra $T(V)=\bigoplus_{n\ge 0}V^{\otimes n}$ satisfies

$$T(V)\cong U(L(V))\cong S(L(V))\cong \bigotimes_{n\ge 1} S(L_n(V)),$$

where $L(V)=\bigoplus_{n\ge 1}L_n(V)$ is the free Lie algebra and $L_n(V)=L(V)\cap V^{\otimes n}$. If $\lambda=(1^{m_1}2^{m_2}\cdots)$, the higher Lie module is

$$L_\lambda(V)=S^{m_1}(L_1(V))\otimes S^{m_2}(L_2(V))\otimes \cdots,$$

and

$T(V)\cong \bigoplus_{\lambda} L_\lambda(V).$

Thrall’s problem asks for the irreducible decomposition of each $L_\lambda(V)$, equivalently the Schur expansion of its character. In the symmetric-group formulation, the same objects appear as induced characters from centralizers of permutations of cycle type $V=\mathbb C^N$0, and the literature alternates among the notations $V=\mathbb C^N$1, $V=\mathbb C^N$2, $V=\mathbb C^N$3, and $V=\mathbb C^N$4 for closely related realizations of the same family (Armon et al., 2024, Adin et al., 16 Sep 2025, Aguiar et al., 13 Aug 2025).

1. Classical definition and equivalent realizations

The classical higher Lie modules are indexed by partitions $V=\mathbb C^N$5, with

$V=\mathbb C^N$6

so that $V=\mathbb C^N$7. In this form, Thrall’s problem is to determine the multiplicity of the irreducible polynomial $V=\mathbb C^N$8-module $V=\mathbb C^N$9 inside $T(V)=\bigoplus_{n\ge 0}V^{\otimes n}$0, ideally by a direct combinatorial count (Ahlbach et al., 2018).

A parallel representation-theoretic model is formulated in terms of centralizers. For $T(V)=\bigoplus_{n\ge 0}V^{\otimes n}$1, if $T(V)=\bigoplus_{n\ge 0}V^{\otimes n}$2 is the centralizer in $T(V)=\bigoplus_{n\ge 0}V^{\otimes n}$3 of a permutation of cycle type $T(V)=\bigoplus_{n\ge 0}V^{\otimes n}$4, then

$T(V)=\bigoplus_{n\ge 0}V^{\otimes n}$5

Choosing a primitive irreducible character $T(V)=\bigoplus_{n\ge 0}V^{\otimes n}$6 of $T(V)=\bigoplus_{n\ge 0}V^{\otimes n}$7, one defines

$T(V)=\bigoplus_{n\ge 0}V^{\otimes n}$8

An equivalent wreath-product recursion is also available: $T(V)=\bigoplus_{n\ge 0}V^{\otimes n}$9 In this language, the irreducible multiplicities are

$$T(V)\cong U(L(V))\cong S(L(V))\cong \bigotimes_{n\ge 1} S(L_n(V)),$$0

and Thrall’s problem becomes the search for a combinatorial interpretation of $$T(V)\cong U(L(V))\cong S(L(V))\cong \bigotimes_{n\ge 1} S(L_n(V)),$$1 (Adin et al., 16 Sep 2025).

The conjugacy-class model is closely related. If $$T(V)\cong U(L(V))\cong S(L(V))\cong \bigotimes_{n\ge 1} S(L_n(V)),$$2 is the conjugacy class of cycle type $$T(V)\cong U(L(V))\cong S(L(V))\cong \bigotimes_{n\ge 1} S(L_n(V)),$$3, and $$T(V)\cong U(L(V))\cong S(L(V))\cong \bigotimes_{n\ge 1} S(L_n(V)),$$4 is the product of the primitive linear characters on the cyclic wreath-product factors of $$T(V)\cong U(L(V))\cong S(L(V))\cong \bigotimes_{n\ge 1} S(L_n(V)),$$5, then

$$T(V)\cong U(L(V))\cong S(L(V))\cong \bigotimes_{n\ge 1} S(L_n(V)),$$6

is a higher Lie character, and the Gessel–Reutenauer theorem identifies the quasisymmetric descent generating function of $$T(V)\cong U(L(V))\cong S(L(V))\cong \bigotimes_{n\ge 1} S(L_n(V)),$$7 with its Frobenius characteristic: $$T(V)\cong U(L(V))\cong S(L(V))\cong \bigotimes_{n\ge 1} S(L_n(V)),$$8 This identification is one of the mechanisms by which higher Lie characters enter the combinatorics of descents and conjugacy classes (Adin et al., 2019).

2. The single-row case and the classical pillars

The case $$T(V)\cong U(L(V))\cong S(L(V))\cong \bigotimes_{n\ge 1} S(L_n(V)),$$9, usually written $L(V)=\bigoplus_{n\ge 1}L_n(V)$0, is the foundational special case. Three classical formulas organize the subject. Klyachko identifies the Schur–Weyl dual of $L(V)=\bigoplus_{n\ge 1}L_n(V)$1 by

$L(V)=\bigoplus_{n\ge 1}L_n(V)$2

Brandt’s formula gives the same character in the power-sum basis: $L(V)=\bigoplus_{n\ge 1}L_n(V)$3 Kraskiewicz–Weyman then gives the Schur coefficients by major-index congruences: $L(V)=\bigoplus_{n\ge 1}L_n(V)$4 This is the cleanest explicit decomposition presently available in the classical theory and remains the prototype for later extensions (Armon et al., 2024).

The same case admits a uniform word-and-necklace interpretation. For the induced cyclic representation $L(V)=\bigoplus_{n\ge 1}L_n(V)$5,

$L(V)=\bigoplus_{n\ge 1}L_n(V)$6

where $L(V)=\bigoplus_{n\ge 1}L_n(V)$7 consists of necklaces of length $L(V)=\bigoplus_{n\ge 1}L_n(V)$8 with frequency dividing $L(V)=\bigoplus_{n\ge 1}L_n(V)$9, $L_n(V)=L(V)\cap V^{\otimes n}$0 consists of words with $L_n(V)=L(V)\cap V^{\otimes n}$1, and $L_n(V)=L(V)\cap V^{\otimes n}$2 consists of words with $L_n(V)=L(V)\cap V^{\otimes n}$3. The proof strategy combines a necklace basis for induced cyclic representations, a content-preserving bijection between $L_n(V)=L(V)\cap V^{\otimes n}$4 and $L_n(V)=L(V)\cap V^{\otimes n}$5, cyclic sieving to exchange $L_n(V)=L(V)\cap V^{\otimes n}$6 with $L_n(V)=L(V)\cap V^{\otimes n}$7, and RSK to pass from word generating functions to Schur expansions (Ahlbach et al., 2018).

These formulas establish the methodological template for the broader subject. The one-row higher Lie character is simultaneously a free-Lie object, an induced cyclic character, a multiplicity-free power-sum expression, and a tableau-counting problem controlled by major index. Much of the later literature can be read as an attempt to preserve these four features for more general $L_n(V)=L(V)\cap V^{\otimes n}$8.

3. Higher Lie modules, rectangles, and induced-character structure

For general $L_n(V)=L(V)\cap V^{\otimes n}$9, one basic reduction is to rectangles. By Littlewood–Richardson theory, it suffices in many constructions to focus on $\lambda=(1^{m_1}2^{m_2}\cdots)$0, since

$\lambda=(1^{m_1}2^{m_2}\cdots)$1

The Schur–Weyl dual of $\lambda=(1^{m_1}2^{m_2}\cdots)$2 is

$\lambda=(1^{m_1}2^{m_2}\cdots)$3

and, more generally, for $\lambda=(1^{m_1}2^{m_2}\cdots)$4, the dual of $\lambda=(1^{m_1}2^{m_2}\cdots)$5 is induced from the centralizer $\lambda=(1^{m_1}2^{m_2}\cdots)$6: $\lambda=(1^{m_1}2^{m_2}\cdots)$7 This makes the higher Lie characters natural examples of induced characters from cyclic or wreath-product centralizers, and it explains why branching rules for $\lambda=(1^{m_1}2^{m_2}\cdots)$8 are structurally central to Thrall’s problem (Ahlbach et al., 2018).

A prime-power deformation sharpens this viewpoint. For each subset $\lambda=(1^{m_1}2^{m_2}\cdots)$9 of the set of primes, the family $$L_\lambda(V)=S^{m_1}(L_1(V))\otimes S^{m_2}(L_2(V))\otimes \cdots,$$0 interpolates between the classical Lie representation and the conjugacy action on $$L_\lambda(V)=S^{m_1}(L_1(V))\otimes S^{m_2}(L_2(V))\otimes \cdots,$$1-cycles: $$L_\lambda(V)=S^{m_1}(L_1(V))\otimes S^{m_2}(L_2(V))\otimes \cdots,$$2 The associated series

$$L_\lambda(V)=S^{m_1}(L_1(V))\otimes S^{m_2}(L_2(V))\otimes \cdots,$$3

has symmetric and exterior powers that serve as analogues of Thrall’s higher Lie modules. The central product formula is

$$L_\lambda(V)=S^{m_1}(L_1(V))\otimes S^{m_2}(L_2(V))\otimes \cdots,$$4

which implies that the coefficient of $$L_\lambda(V)=S^{m_1}(L_1(V))\otimes S^{m_2}(L_2(V))\otimes \cdots,$$5 is a multiplicity-free sum of power sums over partitions whose parts lie in $$L_\lambda(V)=S^{m_1}(L_1(V))\otimes S^{m_2}(L_2(V))\otimes \cdots,$$6. More generally, for a nonempty subset $$L_\lambda(V)=S^{m_1}(L_1(V))\otimes S^{m_2}(L_2(V))\otimes \cdots,$$7, arithmetic Möbius inversion defines

$$L_\lambda(V)=S^{m_1}(L_1(V))\otimes S^{m_2}(L_2(V))\otimes \cdots,$$8

and the higher-Lie-type identity becomes

$$L_\lambda(V)=S^{m_1}(L_1(V))\otimes S^{m_2}(L_2(V))\otimes \cdots,$$9

In plethystic form,

$T(V)\cong \bigoplus_{\lambda} L_\lambda(V).$0

This generalizes Thrall’s construction from one family of Lie characters to a class of symmetric-function systems governed by chosen sets of allowed part sizes (Sundaram, 2021).

4. Tableau formulas and exact solutions for special families

Recent work has made the combinatorial side of Thrall’s problem substantially more explicit. For $T(V)\cong \bigoplus_{\lambda} L_\lambda(V).$1, define consecutive blocks

$T(V)\cong \bigoplus_{\lambda} L_\lambda(V).$2

For $T(V)\cong \bigoplus_{\lambda} L_\lambda(V).$3, the $T(V)\cong \bigoplus_{\lambda} L_\lambda(V).$4-th block descent set and block-major index are

$T(V)\cong \bigoplus_{\lambda} L_\lambda(V).$5

$T(V)\cong \bigoplus_{\lambda} L_\lambda(V).$6

This leads to

$T(V)\cong \bigoplus_{\lambda} L_\lambda(V).$7

A key lemma identifies these block-major indices with major indices of rectified subtableaux, which makes them compatible with product formulas for higher Lie modules (Huh et al., 18 May 2026).

When $T(V)\cong \bigoplus_{\lambda} L_\lambda(V).$8 has distinct parts, the refined tableau condition already gives the exact Schur coefficients: $T(V)\cong \bigoplus_{\lambda} L_\lambda(V).$9 For hook shapes $L_\lambda(V)$0 with $L_\lambda(V)$1,

$L_\lambda(V)$2

These formulas recover the one-row case and extend it to a broader class of partitions (Huh et al., 18 May 2026).

The two-row rectangle $L_\lambda(V)$3 is now solved. For $L_\lambda(V)$4, define

$L_\lambda(V)$5

and

$L_\lambda(V)$6

Then

$L_\lambda(V)$7

The spin statistic is transported from Yamanouchi domino tableaux via a bijection assembled from correspondences of van Leeuwen and the Carré–Leclerc formula for $L_\lambda(V)$8. The same framework extends to all partitions in which every part greater than $L_\lambda(V)$9 occurs at most twice (Huh et al., 18 May 2026).

This line of work suggests a change in the status of the problem. Thrall’s problem remains open in general, but it is no longer confined to isolated one-row formulas: distinct parts, hooks, two rows, and all partitions with each part $V=\mathbb C^N$00 appearing at most twice now admit explicit tableau models (Huh et al., 18 May 2026).

5. Super analogues and extensions beyond type $V=\mathbb C^N$01

A super version of Thrall’s problem starts with a super vector space

$V=\mathbb C^N$02

The free Lie superalgebra inside the tensor superalgebra yields a canonical decomposition

$V=\mathbb C^N$03

where $V=\mathbb C^N$04 is finitely supported with $V=\mathbb C^N$05, and

$V=\mathbb C^N$06

The corresponding super Thrall problem asks for the multiplicity of $V=\mathbb C^N$07 in $V=\mathbb C^N$08 in the stable limit $V=\mathbb C^N$09. In this setting, Brandt’s formula, Klyachko’s theorem, and the Kraskiewicz–Weyman major-index formula all admit super analogues (Armon et al., 2024).

The combinatorial innovation is the introduction of standard super tableaux $V=\mathbb C^N$10, in which each entry $V=\mathbb C^N$11 appears as either $V=\mathbb C^N$12 or $V=\mathbb C^N$13. If $V=\mathbb C^N$14 is the underlying unbarred tableau, the super descent condition is

$V=\mathbb C^N$15

and the super major index is

$V=\mathbb C^N$16

This statistic is designed so that principal specializations of super Schur functions become generating functions in $V=\mathbb C^N$17 and $V=\mathbb C^N$18, and it is characterized by the super $V=\mathbb C^N$19-hook formula

$V=\mathbb C^N$20

The multiplicity formula for the basic super Lie piece is

$V=\mathbb C^N$21

This places the super major index in exactly the role occupied by ordinary major index in the classical one-row theory (Armon et al., 2024).

Higher Lie characters also extend beyond symmetric groups. In type $V=\mathbb C^N$22, for signed cycle type $V=\mathbb C^N$23, centralizers decompose as products of wreath products built from $V=\mathbb C^N$24 and $V=\mathbb C^N$25. Primitive linear characters on these factors define higher Lie characters $V=\mathbb C^N$26, and the $V=\mathbb C^N$27-th root enumerator satisfies

$V=\mathbb C^N$28

In type $V=\mathbb C^N$29, a direct full theory does not always exist, but a restriction-from-$V=\mathbb C^N$30 construction is available in the clean cases, and it is sufficient to prove that $V=\mathbb C^N$31 is a proper character for every integer $V=\mathbb C^N$32 (Adin et al., 2023).

6. Applications, asymptotics, and open directions

Higher Lie characters now appear in several contexts that are not simply reformulations of the original decomposition problem. For conjugacy classes $V=\mathbb C^N$33, the existence of a cyclic descent extension is controlled by hook constituents of the associated higher Lie character. If

$V=\mathbb C^N$34

then $V=\mathbb C^N$35 is divisible by $V=\mathbb C^N$36 if and only if $V=\mathbb C^N$37 is not of the form $V=\mathbb C^N$38 for square-free $V=\mathbb C^N$39. Consequently, the descent map on the conjugacy class $V=\mathbb C^N$40 has a cyclic extension $V=\mathbb C^N$41 if and only if $V=\mathbb C^N$42 is not of the form $V=\mathbb C^N$43 for some square-free integer $V=\mathbb C^N$44. In this application, higher Lie characters serve as the representation-theoretic object from which the decisive hook-multiplicity polynomial is extracted (Adin et al., 2019).

They also admit topological realizations. In the peak-algebra setting, if $V=\mathbb C^N$45 denotes the peak idempotent, then for $V=\mathbb C^N$46 even,

$V=\mathbb C^N$47

where $V=\mathbb C^N$48. Thus peak representations are direct sums of Thrall’s higher Lie characters indexed by the number of odd parts of $V=\mathbb C^N$49, rather than by the number of parts $V=\mathbb C^N$50. The corresponding equivariant Hilbert series is

$V=\mathbb C^N$51

This gives a cohomological model for a nonclassical organization of the higher Lie characters (Aguiar et al., 13 Aug 2025).

From an asymptotic viewpoint, many higher Lie characters become proportional to the regular character. If $V=\mathbb C^N$52 is the higher Lie character for $V=\mathbb C^N$53, then

$V=\mathbb C^N$54

and “tending to be regular” means that for Plancherel-typical irreducibles $V=\mathbb C^N$55,

$V=\mathbb C^N$56

This has been proved for rectangles $V=\mathbb C^N$57 with $V=\mathbb C^N$58 and $V=\mathbb C^N$59, for distinct-row families satisfying explicit growth conditions, and consequently for hooks $V=\mathbb C^N$60 with $V=\mathbb C^N$61. If $V=\mathbb C^N$62 is random with probability

$V=\mathbb C^N$63

equivalently if $V=\mathbb C^N$64 is the cycle type of a uniformly random permutation, then the random higher Lie character tends in probability to be regular (Adin et al., 16 Sep 2025).

The general decomposition problem nonetheless remains open. Several papers now isolate its likely structural ingredients: cyclic induction, necklace generating functions, cyclic sieving, RSK, blockwise major-index congruences, and, in higher-rank settings, new statistics such as the proposed $V=\mathbb C^N$65 that would simultaneously preserve content-class equidistribution and depend only on the recording tableau under RSK. This suggests that the unresolved part of Thrall’s problem is not the absence of structure, but the absence of a single combinatorial statistic with the same naturality for general $V=\mathbb C^N$66 that ordinary major index already has for $V=\mathbb C^N$67 and super major index has for $V=\mathbb C^N$68 in the super setting (Ahlbach et al., 2018, Armon et al., 2024).