REVIEW 4 major objections 4 minor 2 cited by
Character values at elements of order 2
T0 review · 4 major / 4 minor · reviewed 2026-08-11 · deepseek-v4-flash
Pith's one-line read Order-2 elements are singular, yet this paper proves that the character of any highest-weight representation of a classical group or G₂ at such an element is always zero or, up to sign and an explicit power of 2, a product or alternating…
desk verdict Theorem 4.1 for GL(n) is a genuine, fully-proven new result, but the paper's other pillars—Sp/SO by one-sentence analogy and a G2 theorem that contradicts its own proof—are not yet standing. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The machinery is the Weyl character formula taken as a limit: an order-2 element is singular, so the Weyl denominator vanishes, and $\Theta_\lambda$ is obtained as the $\varepsilon \to 0$ limit of the ratio at the nearby regular element $C_{n-k,k}(\varepsilon)$ with coordinates $x_j(\varepsilon) = 1 + j\varepsilon$. The parity sets $\eta_i(\lambda) = \{a \in \lambda + \rho : a \equiv i \bmod 2\}$ organise the rows of the Weyl numerator, and a fixed sequence of column operations ($C_{n-k+i} \to C_{n-k+i} - C_{n-2k+i}$, then scaling by $-1/2$) puts that numerator into block form. Three determinant facts decide the outcome: Lemma 1 (expansion into complementary minors with shuffle signs $\varepsilon_S$), Corollary 1 (a zero submatrix with $a+b > n$ forces the determinant to vanish, giving the vanishing theorem), and Corollary 2 (a zero submatrix with $a+b = n$ splits the determinant into two factors, giving the factorization theorem). In the limit, each surviving factor becomes a character of a smaller classical group at the identity — hence a dimension — and the cross-product $\prod_{s,t}(x_s(\varepsilon) + x_t(\varepsilon))$ becomes the stated power of $2$.
What would settle it
Check the unproved symplectic analogy on one small case: for $\mathrm{Sp}(6,\mathbb{C})$ take $n=3$, $k=1$, the order-2 element $D_{2,1} = (1,1,-1,-1,1,1)$, and the weight $\lambda = (2,2,1)$, for which $\lambda + \rho_3 = (5,4,2)$ and $\#\eta_0(\lambda) = 2 = n-k$. Theorem 5.1B predicts $\Theta_\lambda(D_{2,1}) = \pm 2^1 \cdot \dim S\langle(1,1)\rangle_{\mathbb{C}^4} \cdot \dim S[(2)]_{\mathbb{C}^3} = \pm 2 \cdot 5 \cdot 5 = \pm 50$, where $S\langle(1,1)\rangle_{\mathbb{C}^4}$ is the 5-dimensional irreducible of $\mathrm{Sp}(4,\mathbb{C})$ and $S[(2)]_{\mathbb{C}^3}$ the 5-dimensional irreducible of $\mathrm{SO}(3,\mathbb{C})$. Evaluating $\Theta_\lambda$ at $D_{2,1}$ directly (via the limiting Weyl character formula of the paper, or with a computer algebra system) and comparing with $\pm 50$ settles whether the stated constant $2^{(n-2k)^2}$ is correct.
Extended reading notes
Core claim
The central claim, proved in detail for $\mathrm{GL}(n,\mathbb{C})$ (Theorem 4.1) and stated by analogy for the other groups, is a complete description of order-2 character values. Fix the involution $C_{n-k,k} = (1,\ldots,1,-1,\ldots,-1)$ with $n-k$ ones and $k$ minus-ones, and (after a determinant-character twist if needed) suppose $\#\eta_0(\lambda) \geq \#\eta_1(\lambda)$, where $\eta_0(\lambda)$ and $\eta_1(\lambda)$ are the even and odd coordinates of $\lambda + \rho$. Then $\Theta_\lambda(C_{n-k,k})$ vanishes when $\#\eta_0(\lambda) > n-k$; when $\#\eta_0(\lambda) = n-k$ it equals $\pm 2^{c(k)} \dim\bigl(S(\lambda_0)_{\mathbb{C}^{n-k}} \otimes S(\lambda_1)_{\mathbb{C}^k}\bigr)$ with $c(k) = \binom{n-2k}{2}$, where $\lambda_0 + \rho_{n-k} = \eta_0(\lambda)/2$ and $\lambda_1 + \rho_k = [\eta_1(\lambda)-1]/2$; and when $\#\eta_0(\lambda) < n-k$ it is a signed alternating sum, over $k$-subsets of the odd coordinates, of such tensor-product dimensions divided by $2^{k(n-k-1)}$. The symplectic and even-orthogonal analogues (Theorems 5.1 and 6.1) follow the same schema with constants $2^{(n-2k)^2}$ and $2^{2\binom{n-2k}{2}-k+1}$, the latter built from Spin representations of smaller even orthogonal groups, while $\mathrm{SO}(2n+1,\mathbb{C})$ (Theorem 7.1) has no vanishing part. For $\mathrm{G}_2(\mathbb{C})$, the character at the unique order-2 class is zero exactly when $k$ and $l$ are both odd, and otherwise factors as a product of two $\mathrm{SL}_2(\mathbb{C})$ characters evaluated at $x^2$ and $x^3$, specialising at $x=1$ to explicit quadratic polynomials in $k$ and $l$ (Theorem 8.1, Proposition 2).
Load-bearing premise
Theorems 5.1, 6.1, and 7.1 defer their proofs to the statement that the argument is analogous to the proof for $\mathrm{GL}(n,\mathbb{C})$: no column-operation or shuffle-sign computation is written down for the symplectic and orthogonal Weyl numerators, so the stated constants $2^{(n-2k)^2}$, $2^{2\binom{n-2k}{2}-k+1}$, and $2^{2kn-2k^2+k-1}$ all depend on that analogy holding.
Editorial extensions
If this is right
- Every order-2 character value of $\mathrm{GL}(n,\mathbb{C})$ is decided by one parity count: comparing $\#\eta_0(\lambda)$ with $n-k$ selects zero, a two-factor dimension product, or an alternating sum over $\binom{\#\eta_1(\lambda)}{k}$ terms.
- The same parity count governs $\mathrm{Sp}(2n,\mathbb{C})$ and $\mathrm{SO}(2n,\mathbb{C})$, with constants $2^{(n-2k)^2}$ and $2^{2\binom{n-2k}{2}-k+1}$, the latter involving Spin representations of the smaller even orthogonal groups.
- In the balanced cases ($n = 2k$ or $n = 2k+1$ for $\mathrm{GL}$, $n = 2k$ for $\mathrm{Sp}$) the power of $2$ is $1$, so the character value is, up to sign, exactly one dimension, recovering the order-2 case of the factorization theorems of [DP] and [AK].
- For $\mathrm{G}_2(\mathbb{C})$ the order-2 character value is an explicit quadratic polynomial in the highest-weight coefficients $k,l$, vanishing exactly when both are odd.
- $\mathrm{SO}(2n+1,\mathbb{C})$ is the exception: it has no vanishing theorem, and its order-2 values are signed sums (single dimensions only in the $n = 2k$ case) of dimensions of smaller orthogonal-group representations.
Reading between the lines
- The three exponents $c(k) = \binom{n-2k}{2}$, $d(k) = (n-2k)^2$, and $e(k) = 2\binom{n-2k}{2} - k + 1$ probably share one combinatorial meaning — plausibly the count of even-coordinate pairs forced by the limit — that the deferred 'analogous' proofs would expose; if so, the same exponent should reappear as a multiplicity in the restriction of the representation to the involution's fixed subgroup.
- Remark 4 asks what is special about the element $(x,-x,-x^{-2})$ in $\mathrm{SL}_3(\mathbb{C})$; the proof identifies it with the quotient of $\mathrm{SL}_2 \times \mathrm{SL}_2$ by $(-1,-1)$ inside $\mathrm{G}_2$, so a testable extension is that factorization occurs precisely along the image of that homomorphism.
- Since $\mathrm{SO}(2n+1,\mathbb{C})$ has no vanishing theorem, an explicit vanishing criterion for $B_n$ remains open; the natural route is the same $\varepsilon \to 0$ Weyl-limit computation with the half-integer shifts of $\rho_{2n+1}$, which should yield a parity condition involving the extra fixed coordinate $1$.
Signed reviews
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper proposes explicit formulas, in terms of products or alternating sums of dimensions of highest-weight representations of smaller classical groups, for the values of irreducible characters at diagonal elements of order 2 for GL(n,C), Sp(2n,C), SO(2n,C), SO(2n+1,C), and G2(C). The GL(n,C) result is proved in detail from the Weyl character formula, using a determinant expansion and explicit column operations. The symplectic and orthogonal theorems are each dispatched with a one-sentence statement that the proof is analogous, and the odd orthogonal theorem explicitly omits the proofs of both parts. The G2(C) section derives factorizations from the Fulton-Harris SL3(C)-restriction character formula and compares them with a formula attributed to Reeder.
Significance. If the full set of formulas were correct, the paper would give a uniform and attractive reduction of order-2 character values to dimensions of smaller representations, together with parity-based vanishing criteria. The GL(n) theorem appears to be a genuine and checkable result, and the determinant manipulations in Section 4 are explicit enough to verify. However, the advertised scope is far larger than what is actually demonstrated: the C_n, D_n, and B_n assertions are unsupported as submitted, and the G2 theorem is internally inconsistent and false as stated. The significance of the full paper is therefore not established, although the GL(n) portion has independent value.
major comments (4)
- [§8, Theorem 8.1] The proof of Theorem 8.1 states that S(1,1,0)(X)-S(1,0,0)(X) = -x^2 - x^{-2} for X=(x,-x,-x^{-2}). Direct substitution gives S(1,1,0)(X) = -x^2 and S(1,0,0)(X) = -x^{-2}, so the difference is -x^2 + x^{-2}. This is not a cosmetic sign error: the evaluation at x=1 is a 0/0 limit, and replacing the denominator by the printed expression changes the limiting value of the quotient.
- [§8, Theorem 8.1, Case II] For k even and l odd, the statement of Theorem 8.1 assigns SL2(C) factors of highest weights (3l+k)/4 and (k+l)/2, whereas the proof of Case II concludes with factors of highest weights (2k+3l+1)/4 and (l-1)/2. For (k,l)=(2,1) these are (5/4,3/2) versus (2,0). Neither 5/4 nor 3/2 is a highest weight of an irreducible algebraic SL2(C) representation, and no definition of characters with half-integral highest weights is supplied. If one formally evaluates the displayed product at x=1 one obtains -45/8, while the proof's own product gives +3 and Proposition 2 gives -3. Thus the theorem and its proof disagree, and the G2 claim is false as written.
- [§5–§7, Theorems 5.1, 6.1, 7.1] The theorems for Sp(2n,C), SO(2n,C), and SO(2n+1,C) are central to the abstract's claim, but no computations are provided for them. Theorem 5.1 and Theorem 6.1 each say only that the proof is analogous to the GL(n) proof, and Theorem 7.1 explicitly says that the proofs of both parts are omitted. In particular, the constants 2^{d(k)}, 2^{e(k)}, and 2^{2kn-2k^2+k-1}, the two-determinant SO(2n) numerator, and the half-integer shifted SO(2n+1) numerator are never derived. The B_n, C_n, and D_n results are therefore unsupported as submitted.
- [§8, Proposition 2] The sentence introducing Proposition 2 states that its proof is a direct consequence of Theorem 5.1 evaluated at (x,-x,-x^{-2}) for x=1. This cannot be correct as printed, because Theorem 5.1 concerns Sp(2n,C) and its element D_{n-k,k}, not G2(C). Since Proposition 2 is used to interpret the G2 character values in Remark 5, its actual source (Reeder [R]) and its logical relation to Theorem 8.1 should be stated correctly.
minor comments (4)
- [§7, Theorem 7.1(A)] The text says that the highest weight λ0 of GL(2m,C) is given by the recipe of Theorem 2.17 in [AK], but m is not defined; in this part n=2k, so the group should presumably be GL(2k,C) or GL(n,C).
- [§7, Theorem 7.1] The notation S[λ]C_{2n} is used for a representation of SO(2n+1,C); the subscript C_{2n} appears to be a typo for C_{2n+1}, since the ambient group has dimension 2n+1.
- [§8, Proposition 1] The proposition uses both Π_{k,l} and Θ_{k,l} for the G2 highest weight representation and its character; the notation should be made consistent.
- [General] The theorem statements rely on the parity sets η_i(λ), but the dependence of the final formulas on the choice of the determinant normalization in the GL(n) case is only described informally; making that normalization explicit would improve readability.
Circularity Check
No significant circularity; the central GL(n) derivation is a self-contained Weyl character formula computation.
full rationale
The main derivation in Section 4 is self-contained: Theorem 4.1 starts from the Weyl character formula, takes a regular perturbation C_{n-k,k}(ε), applies elementary determinant identities (Corollaries 1–2 and Lemma 1), evaluates the limit, and obtains the claimed factorization, vanishing criterion, and alternating sum with the powers of 2 emerging from the denominator. There are no fitted parameters, and the result is not assumed anywhere in the proof. The later theorems for Sp, SO(2n), and SO(2n+1) are asserted by one-sentence analogies or have their proofs omitted (Section 7 explicitly says so), and Theorem 8.1 for G2 is internally inconsistent, with a sign error in the printed denominator and a mismatch between the displayed SL2 weights and the calculation in the proof; however, these are correctness and support defects, not circularity. The citations to [DP], [AK], [FH], and [R] are external sources used as ingredients or as alternate derivations, and there is no self-citation chain that carries the central claim. Proposition 2 is explicitly credited to Reeder, so the paper's reliance on that formula is not a circular self-import.
Assumptions & free parameters
assumptions (5)
- standard math Weyl character formula for GL(n), Sp(2n), SO(2n), SO(2n+1) as a quotient of alternating determinants, valid also in the regular-element limit at singular points
- standard math Lemma 1 and Corollaries 1-2: generalized Laplace expansion expressing det B as a signed sum over k-column submatrices
- domain assumption The listed elements C_{n-k,k}, D_{n-k,k}, E_{n-k,k}, F_{n-k,k} exhaust the conjugacy classes of order 2 in the respective groups
- domain assumption Proposition 1 (Fulton-Harris, Prop 24.48): character of G2 irreps as a Schur-polynomial quotient for the SL3 embedded in G2
- ad hoc to paper The character at x = 1 evaluation of the SL2 factors is well-defined even when intermediate highest weights are half-integers, via the quotient of SL2 x SL2 by (-1,-1)
Cite this review
Pith. "Pith review of Character values at elements of order 2." pith.science (2026). https://pith.science/paper/D4NRYK37
@misc{pith2026241217324,
author = {Pith},
title = {Pith review of: Character values at elements of order 2},
year = {2026},
howpublished = {\url{https://pith.science/paper/D4NRYK37}},
note = {Machine review of arXiv:2412.17324}
}
read the original abstract
In this paper we compute the character values of highest weight representations for classical groups of types A_n, B_n, C_n, D_n and the Exceptional group G_2 at all conjugacy classes of order 2. We prove that these character values, if nonzero, can be expressed either as a product involving the dimensions of two highest weight representations from classical subgroups, along with an additional constant term, or as an alternating sum of products of the dimensions of two highest weight representations from the classical subgroups, also accompanied by an extra constant term.
Forward citations
Cited by 2 Pith papers
-
Factorization of Schur polynomials twisted by roots of unity and a reciprocal pair
For s_λ(μ_t, z, z^{-1}), the evaluation is zero or a signed product of three hyperbolic sine factors read from the t-residue profile, for every t and every shape.
-
Character theory at a torsion element
Character values at principal torsion elements of a compact Lie group equal, up to sign and a constant, the dimension of a representation of a dual centralizer group.
Reference graph
Works this paper leans on
-
[1]
W. Fulton and J. Harris, Representation Theory A First Course. Springer-Verlag, New York, 1991
work page 1991
-
[2]
A. Ayyer and N. Kumari, Factorization of Classical characters Twisted by Roots of Unity. J. Algebra 609 (2022), 437–483
work page 2022
-
[3]
B. Kostant. On Macdonald's -function formula, the Laplacian and generalized exponents. Adv. Math. 20 (1976), no. 2, 179–212
work page 1976
-
[4]
Transformation Groups 29 (2024), no.3, 1161-1198
Reeder, Mark, Weyl Group Characters Afforded By Zero Weight Spaces. Transformation Groups 29 (2024), no.3, 1161-1198
work page 2024
-
[5]
Prasad, A character relationship on GL (n, C )
D. Prasad, A character relationship on GL (n, C ) . Israel J. Math. 211 (2016), no. 1, 257–270
work page 2016
-
[6]
Kumari, Nishu, Factorization of Classical characters Twisted by Roots of Unity: II. J. Pure Appl. Algebra 228 (2024), no. 11, Paper No. 107714, 45 pp
work page 2024
Reviewed August 11, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.