REVIEW 4 minor 2 references
Higher signs for Coxeter groups
T0 review · 0 major / 4 minor · reviewed 2026-08-14 · deepseek-v4-flash
Pith's one-line read Every Coxeter group carries a uniquely determined 'higher sign' in every degree, and for Weyl groups the degree-three sign is forced by the associativity of convolution in geometric categories.
desk verdict Strong, self-contained construction of higher sign cocycles for all Coxeter groups, with a clean uniqueness theorem; the geometric comparison is plausible but rests on prior work. 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 universal collapsing cocycle $$Z_n^W: W^n \to \mathbb{Z}[\mathcal{D}]^{\varepsilon(n)},$$ defined by summing, over walls $H$ of the geometric realization, a signed contribution whenever the chain of chambers $(C_0,\ldots,C_n)$ crosses $H$ at every step. The sign is chosen so that $Z_n^W$ lands in the invariant part $\mathbb{Z}[\mathcal{D}]^+$ for even $n$ and the anti-invariant part $\mathbb{Z}[\mathcal{D}]^-$ for odd $n$, matching the parity of the cocycle degree. This cocycle is universal among collapsing $n$-cocycles with $\mathbb{Z}[W]$-module coefficients: a $W$-equivariant homomorphism $\mathbb{Z}[\mathcal{D}]^{\varepsilon(n)}\to A$ sends $Z_n^W$ to any given collapsing cocycle $\zeta$. The higher signs $\varepsilon_n^W$ are the images under the coefficient homomorphism $\chi_n$, so they count walls crossed by an alternating chamber chain, with parity when $n$ is odd. For $n=3$ in the Weyl-group case, the geometric identification uses the groupoid $\Xi$ of blocks in monodromic Hecke categories—categories of sheaves on a reductive group equivariant under a torus action—and the canonical isomorphisms of convolution; the resulting cocycle $\sigma$ is collapsing and matches $\varepsilon_3^W$ after pulling back along the map from blocks to group elements.
What would settle it
Compute $\varepsilon_3^W$ for $W=S_3$ on the triple $(s,t,s)$ of the two simple reflections; the theorem forces the value $0$ by collapsing, and Theorem 4.4 forces the corresponding geometric associativity sign for convolution of perverse sheaves on the flag variety to be $+1$. A direct geometric calculation producing $-1$ on this triple—or on any triple where $\varepsilon_3^W$ is predicted to vanish—would disprove the identification $\sigma=\pi^*\varepsilon_3^W$.
Extended reading notes
Core claim
The central claim is Theorem 1.1: for any Coxeter group $(W,S)$ and any $n\ge 1$ there is exactly one $n$-cocycle $\varepsilon_n^W$ satisfying two conditions: it collapses, meaning it vanishes on any $n$-tuple where some adjacent pair $x_i,x_{i+1}$ has length $|x_i x_{i+1}|=|x_i|+|x_{i+1}|$, and it evaluates to $1$ on the diagonal tuple $(s,\ldots,s)$ for every simple reflection $s$. Equivalently, the cocycle $Z_n^W$ built from the geometric realization—counting walls separated by an alternating chain of chambers with signs—is the universal collapsing $n$-cocycle: every collapsing cocycle with values in a $\mathbb{Z}[W]$-module factors through it uniquely. The higher signs $\varepsilon_n^W$ are the images of $Z_n^W$ under the coefficient map sending each wall to the same value, so they count (with parity when $n$ is odd) the walls crossed by an alternating chamber chain. For Weyl groups, Theorem 4.4 identifies $\varepsilon_3^W$ with the pullback of the $3$-cocycle $\sigma$ on the groupoid of blocks of monodromic Hecke categories, so the degree-three sign is the associativity constraint of convolution.
Load-bearing premise
The geometric half of the paper rests on a large prior construction: if the block decomposition of the relevant sheaf categories, or the explicit isomorphisms defining the geometric cocycle, are incorrect, the claimed equality between the combinatorial and geometric three-cocycles could fail, even though the combinatorial cocycles stand on their own.
Editorial extensions
If this is right
- The usual sign homomorphism is exactly the $n=1$ case, and the $n=2$ case recovers the integer-valued two-cocycle $(x,y)\mapsto \tfrac12(|x|+|y|-|xy|)$.
- Every collapsing $n$-cocycle of $W$ with $\mathbb{Z}[W]$-module coefficients is a unique pushforward of $Z_n^W$, so computations with collapsing cocycles reduce to chamber geometry.
- For Weyl groups, the degree-three cocycle is forced by the associativity constraint of convolution in monodromic Hecke categories; any monoidal structure on those categories must carry the sign $\varepsilon_3^W$.
- The restrictions of $\varepsilon_3^W$ to the stabilizer $\Omega$ of the fundamental alcove are often nontrivial in cohomology: explicit classes are computed for type $A_n$ with $n$ even, $B_n$, $C_n$, $D_n$, and $E_7$, showing that the twisting is genuinely nontrivial.
- The higher signs are compatible with automorphisms, parabolic subgroups, and direct products, and they satisfy inversion symmetry; hence they define canonical invariants of the Coxeter group rather than artifacts of a presentation.
Reading between the lines
- Editorial inference: Because $Z_n^W$ is universal, the same wall-counting construction should yield collapsing cocycles for any group acting on a chamber complex or a building, not only Coxeter groups.
- Editorial inference: The remark that $\varepsilon_n^W$ is not a cup power of the sign suggests that the classes $[\varepsilon_n^W]$ for $n\ge 3$ are genuinely new cohomology classes, potentially encoding connectivity or curvature data of the Coxeter complex.
- Editorial inference: The nontrivial restrictions computed in Section 5 could serve as a diagnostic for whether a given block decomposition of Hecke categories carries nontrivial monoidal twisting, with possible consequences for classifying indecomposable objects or computing extension groups.
- Editorial inference: The degree-three cocycle's appearance as an associativity constraint suggests that higher signs may appear as higher associativity constraints in $n$-category or $A_\infty$ enrichments of Hecke categories, though the paper explicitly leaves this open for $n\ge 4$.
Signed reviews
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper defines, for each Coxeter group (W,S) and each positive integer n, a canonical n-cocycle epsilon_n^W taking values in Z for even n and in F_2 for odd n. The construction is via a universal collapsing n-cocycle Z_n^W built from alternating chamber-wall data in the geometric realization of W. Theorem 1.1 states uniqueness of epsilon_n^W under two conditions: vanishing whenever some adjacent pair of arguments multiplies without length cancellation, and value 1 on the diagonal (s,...,s). The paper also gives explicit formulas, relates epsilon_3 to a geometric 3-cocycle in monodromic Hecke categories for Weyl groups, and computes restrictions of epsilon_3 to the finite abelian group Omega attached to extended affine Weyl groups in types A, B, C, D, and E_7.
Significance. The main combinatorial result, Theorem 1.1 together with the universal property in Theorem 2.9, is a clean and convincing generalization of the sign character, and the proof in Sections 2 and 3 is detailed and essentially self-contained. The wall-and-chamber formula gives an explicit, computable cocycle, and the paper derives several useful structural properties: the n=2 case recovers the Tits extension data, odd-degree cocycles are related to even-degree ones by the Bockstein homomorphism, and the restriction computations in Section 5 give concrete nontriviality statements. If the geometric comparison in Theorem 4.4 is accepted, the paper also provides a natural representation-theoretic origin for epsilon_3; however, that portion is conditional on the companion paper [2]. The central existence and uniqueness of the higher signs does not depend on [2] and appears sound.
minor comments (4)
- [2.11] In the proof of Lemma 2.11, the sentence ending with 'by (2.10)' refers to a nonexistent equation; it should refer to Lemma 2.10.
- [2.11] In the base case L(x)=n of Lemma 2.11, the argument implicitly uses that no entry can have length at least 2: if all n entries have positive length and their sum is n, then all lengths must be 1. This one-line justification should be stated explicitly.
- [4.3-4.4] Lemma 4.3 is only a proof sketch and Theorem 4.4 depends on it together with results from [2] such as Lemma 4.5(3)(4) and Section 5.3; the authors should state clearly in the text that the geometric comparison is conditional on the companion paper, or expand the proof of Lemma 4.3.
- [2.5] In the definition of Z_n^W in (2.1), the chamber C_0 is used but only C_i for i>=1 is defined; a sentence fixing C_0=C would remove a small ambiguity.
Circularity Check
No significant circularity: the construction and uniqueness of ε_n^W are self-contained; the geometric comparison in Section 4 depends on prior work [2] but does not reduce to it.
full rationale
The central derivation in Sections 2–3 is self-contained. The paper constructs Z_n^W directly from chamber/wall combinatorics, proves directly that it is a collapsing cocycle (Proposition 2.7), proves that it is universal among collapsing cocycles by constructing an explicit W-equivariant map from its values (Theorem 2.9), and then defines ε_n^W as the image under the invariant map χ_n. The uniqueness claim in Theorem 1.1 is proved from Lemma 2.11, which shows that any collapsing cocycle vanishing on (s, ..., s) is identically zero; this lemma does not presuppose the existence or values of ε_n^W. The geometric comparison in Section 4 relies on the prior article [2] for the definition of the monodromic Hecke category, the groupoid Ξ, the canonical isomorphisms, and the SL2 base value σ(β, β^{-1}, β) = -1. However, [2] is used as a source of technical geometric inputs, not as a source of the target cocycle ε_3^W or its properties. Theorem 4.4 proves σ = π^*ε_3^W by checking that σ is collapsing and then applying the vanishing lemma, so the cited results from [2] are independent inputs rather than restatements or fitted versions of the conclusion. No fitted parameter is renamed as a prediction, no uniqueness theorem is imported from the authors to force the choice, and no ansatz is smuggled in via citation. The dependence of Theorem 4.4 on [2] is a matter of external verification of prior work, not circularity.
Assumptions & free parameters
assumptions (5)
- standard math Geometric realization of a Coxeter group by reflections: there is a faithful reflection representation on a real vector space with walls, chambers, and the set of reflecting half-spaces (Bourbaki, Ch. V §4).
- standard math Standard definitions and properties of group cohomology (cocycles, coboundaries, Bockstein homomorphism, connecting homomorphisms).
- domain assumption Existence and properties of the Tits section for Weyl groups (Example 2.2).
- domain assumption All results of [2] (Lusztig-Yun) used in Section 4: monodromic Hecke category blocks, groupoid Xi, convolution isomorphisms, the 3-cocycle sigma, reduced decomposition of morphisms, and affine space bundle facts.
- standard math Facts about perverse sheaf convolution, extensions, and cohomology of affine spaces and Gm used in Section 4.4 and Remark 4.5.
Cite this review
Pith. "Pith review of Higher signs for Coxeter groups." pith.science (2026). https://pith.science/paper/M4BDX4NK
@misc{pith2026190803672,
author = {Pith},
title = {Pith review of: Higher signs for Coxeter groups},
year = {2026},
howpublished = {\url{https://pith.science/paper/M4BDX4NK}},
note = {Machine review of arXiv:1908.03672}
}
read the original abstract
We define and study cocycles on a Coxeter group in each degree generalizing the sign function. When the Coxeter group is a Weyl group, we explain how the degree three cocycle arises naturally from geometry representation theory.
Reference graph
Works this paper leans on
-
[2]
Endoscopy for Hecke categories and c haracter sheaves
Lusztig, G.; Yun,Z. Endoscopy for Hecke categories and c haracter sheaves. arXiv:1904.01176. Department of Mathematics, Massachusetts Institute of Tec hnology, 77 Massachusetts A ve, Cambridge, MA 02139 E-mail address : zyun@mit.edu
arXiv 1904
-
[1]
Bourbaki, N. ´El´ ements de math´ ematique. Fasc. XXXIV. Groupes et alg` ebres de Lie. Chapitre IV-VI. Actualit´ es Scien- tifiques et Industrielles, No. 1337 Hermann, Paris 1968, 288 pp
work page 1968
Reviewed August 14, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.