REVIEW 1 major objections 5 minor
Comparing the face rings of a boolean complex and its barycentric subdivision
T0 review · 1 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash
Pith's one-line read For a Cohen–Macaulay complex, the face ring and its barycentric subdivision ring are equivariantly isomorphic over the parameter subring when the automorphism-group order is invertible in the field; a simplex in characteristic 2 fails.
desk verdict Positive theorem is solid; negative theorem hinges on a compressed Lemma 5.4 that deserves verification before publication. 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 load-bearing object is the Garsia transfer $G\colon k[\operatorname{Sd}\Delta]\to k[\Delta]$, which sends each generator $y_\alpha$ to $x_\alpha$ and extends multiplicatively on standard monomials; it is a graded, $\operatorname{Aut}(\Delta)$-equivariant $k$-linear isomorphism that need not be a ring or module map. Its power comes from the shape grading of $k[\operatorname{Sd}\Delta]$ by partitions and the compatible shape filtration of $k[\Delta]$ by dominance order (Lemma 3.17 and Proposition 3.18), under which the transfer is a 'homomorphism in the top shape': products of standard monomials move to shapes no larger than the sum of the factors, with equality only when the factors stack up. Theorem 3.28 uses this to transfer generation and linear independence from $k[\operatorname{Sd}\Delta]$ to $k[\Delta]$, and a separate averaging lemma (Proposition 6.10) converts any shape-filtered isomorphism that is equivariant in the top shape into a fully $G$-equivariant isomorphism when $|G|$ is invertible in $k$. Section 4's subspace-arrangement characterization of Cohen–Macaulayness drives Algorithm 6.11, which constructs the required shape-homogeneous basis.
What would settle it
Carry out a direct Gröbner-basis or module computation over $\mathbb{F}_2$ of the $A_n$-invariants of $k[\operatorname{Sd}\Delta_d]$ for the smallest simplex $d=2$ ($n=3$): if the invariants are not a free rank-two $k[\Theta]$-module with basis $\{1,\tilde D\}$, Lemma 5.4 is false and Theorem 1.2 loses its foundation; if they are, the counterexample stands. Running Algorithm 6.11 on a balanced boolean complex over a field whose characteristic divides $|\operatorname{Aut}(\Delta)|$ would additionally test whether the coprime hypothesis marks the exact boundary of the positive result.
Extended reading notes
Core claim
The paper's central claim is Theorem 1.3: for a finite boolean complex $\Delta$ that is Cohen–Macaulay over $k$, and a group $G$ of automorphisms whose order is a unit in $k$, there exists a graded, $G$-equivariant isomorphism $k[\operatorname{Sd}\Delta]\to k[\Delta]$ of modules over the common parameter subring $k[\Theta]$, together with an algorithm that computes it. The complementary Theorem 1.2 says that for a $d$-simplex with $d\ge 2$ over a field of characteristic $2$, no such equivariant isomorphism exists for the full automorphism group $S_n$; this shows the equivariant version of Murai's question fails without the coprime hypothesis. The positive construction takes a shape-homogeneous basis of $k[\operatorname{Sd}\Delta]$ over $k[\Theta]$, transfers it by the Garsia map to a basis of $k[\Delta]$, forms the resulting module isomorphism, and then averages over $G$, with the shape filtration guaranteeing that averaging remains invertible.
Load-bearing premise
The counterexample rests on the assertion that in characteristic $2$ the alternating-group invariant subring $k[\operatorname{Sd}\Delta]^{A_n}$ is a free module of rank two over the parameter subring with basis $\{1,\tilde D\}$; the paper supplies only a proof sketch for this step, and if the assertion failed, the contradiction behind Theorem 1.2 would collapse.
Editorial extensions
If this is right
- In the Cohen–Macaulay coprime setting, the face ring and the barycentric-subdivision ring carry the same $G$-equivariant structure over the parameter ring, so all equivariant Betti numbers over $k[\Theta]$ coincide.
- The algorithm makes the isomorphism computable from a shape-homogeneous basis; the Gröbner-free linear-algebraic route only requires facet-vector row reductions.
- The simplex counterexample in characteristic $2$ shows the equivariant version fails for a Cohen–Macaulay complex when the characteristic divides $|\operatorname{Aut}(\Delta)|$, so the unit-order hypothesis in Theorem 1.3 cannot simply be dropped.
- The non-equivariant answer to Murai's question and Adams–Reiner's original Betti-number conjecture remain plausibly true in full generality, as the paper notes.
Reading between the lines
- The averaging deformation is phrased for barycentric subdivisions, but the same argument would turn any shape-filtered, top-shape-equivariant module isomorphism between two graded $G$-rings with equal Hilbert series into a fully equivariant one whenever $|G|$ is invertible.
- For a fixed complex whose full automorphism group has order divisible by the characteristic, the positive theorem still applies to each subgroup of order coprime to the characteristic, so the symmetry that survives is precisely the symmetry whose order is invertible in the field.
- Algorithm 6.11 could serve as a combinatorial Cohen–Macaulayness test for balanced boolean complexes: its termination behavior gives a certificate, which may be implemented and compared with the homology criterion of Reisner–Munkres.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper studies Murai's question whether the Stanley–Reisner ring (face ring) of a finite boolean complex and that of its barycentric subdivision are isomorphic as modules over the common parameter subring k[Θ], and whether such an isomorphism can be chosen equivariantly with respect to a group of automorphisms. The main positive result, Theorem 1.3, states that if Δ is Cohen–Macaulay over k and |G| is a unit in k for G ⊆ Aut(Δ), then a graded G-equivariant k[Θ]-module isomorphism k[Sd Δ] → k[Δ] exists and can be computed by an explicit algorithm. The main negative result, Theorem 1.2, states that for the d-simplex with d ≥ 2 over a field of characteristic 2, no Aut(Δ)-equivariant k[Θ]-module isomorphism exists. The proofs are built on a detailed generalization of Garsia's transfer method and on a linear-algebraic characterization of Cohen–Macaulayness for balanced boolean complexes.
Significance. If the results stand, the paper gives a definitive answer to Murai's question under the Cohen–Macaulay/coprime hypothesis and, in the modular simplex case, a counterexample. The positive theorem is constructive: it supplies an algorithm, not merely an existence statement, and the reworking of Garsia's tools at the generality of boolean complexes is a substantive contribution in its own right. The nonconstructive proof via equivariant Hilbert series and the constructive transfer argument complement each other nicely. The main caveat is that the negative theorem rests on Lemma 5.4, whose proof is currently only a sketch; the statement is plausible and the gap appears repairable, but the proof as written is incomplete.
major comments (1)
- [§5, Lemma 5.4] Lemma 5.4 is the only load-bearing step in the proof of Theorem 1.2 that is not proved in the text. The lemma asserts that in characteristic 2, k[Sd Δ]^{A_n} is a free k[Θ]-module of rank two with basis {1, \tilde D}; the proof is a sketch citing [Rei92, Theorem 4.3.5] and [GS84, Theorem 6.2], with the assertion that the characteristic-zero hypothesis of the latter is not required. The subsequent contradiction depends literally on this statement: the proof writes φ(\tilde D) = u + vD and uses that representation, together with the freeness and basis, to derive equation (16). If the invariant subring is not free of rank two with the stated basis, or if \tilde D is not a basis element, the divisibility argument collapses. The manuscript should give a complete proof of this lemma, or alternatively a precise citation to a statement that explicitly covers the modular case, plus a direct verification in the smallest case d = 2 over F_2. This is essential to make Theorem 1.2 fully supported.
minor comments (5)
- [§6.2] When Theorem 3.28 is used to transfer the basis b_1, ..., b_r in the construction of Φ, the text should say explicitly that the theorem is being applied with the trivial group; otherwise the unstated hypothesis that the elements be G-invariant is puzzling, since the b_j need not be G-invariant.
- [§3.3, proof of Theorem 3.28] In the maximal-shape argument of the proof of assertion 2, after choosing a chain supporting a term of G(f_j), the text should explicitly say that the chain is extended to a maximal chain so that an element of every rank is available; as written, the existence of the required stacking monomial of shape a_1(1^1)+...+a_n(1^n) is not immediate when the original chain misses some ranks.
- [§4, Lemma 4.6] The claim that a standard monomial m not sitting under any facet to which β belongs satisfies m z_β = 0 deserves a one-sentence justification; the point is that any common upper bound of m and β would extend to a facet containing both, contradicting the hypothesis.
- [§3.2, proof of Lemma 3.17] The comparison of shapes in the straightening move uses implicitly the rank identity rk(α)+rk(β)=rk(α∧β)+rk(γ); the text should state this identity explicitly, since otherwise the two shapes being compared are not visibly partitions of the same integer.
- [§5] The notation for the invariant basis element is inconsistent: Lemma 5.4 writes \tilde D while the proof of Theorem 1.2 repeatedly writes pD. Please standardize the notation.
Circularity Check
No circularity: the positive construction and the negative counterexample are derived from explicit bases, transfer arguments, and a genuine divisibility contradiction, with only a non-circular proof gap in Lemma 5.4.
full rationale
The positive result (Theorem 1.3) is not circular. The construction first produces a shape-homogeneous k[Theta]-basis for k[Sd Delta] via Algorithm 6.11 in Section 6.3, transfers it to a k[Theta]-basis of k[Delta] using the Garsia transfer (Theorem 3.28), defines Phi by sending each basis element b_j to G(b_j), proves that Phi is shape-filtered and G-equivariant in the top shape (Lemma 6.8 and Proposition 6.9), and then averages over G in Proposition 6.10 to obtain a fully equivariant isomorphism. Each of these statements is proved in the paper from the shape-filtration and associated-graded mechanism of Section 3, not assumed from the target claim. The nonconstructive existence proof in Section 6.1 similarly derives the isomorphism from equality of equivariant Hilbert series and Lemma 6.2, which is proved independently from semisimplicity of kG in coprime characteristic. The negative result (Theorem 1.2) is also a genuine contradiction rather than a repackaged assumption. Assuming an equivariant isomorphism, the paper restricts to A_n-invariants, writes the restricted map on the explicit rank-two free bases supplied by Lemmas 5.3 and 5.4, derives the equation theta_1...theta_d s = v(D + tau D), and then contradicts it with the explicit cross-term of Lemma 5.5. The contradiction does not presuppose the nonexistence of the isomorphism, and the lemmas are not fitted to the theorem's conclusion. The paper contains several citations to the first author's thesis [BS17], for example in Lemma 3.8, in the remark after Theorem 3.28, in Lemma 5.2, and in the provenance discussion of Algorithm 6.11. These are either accompanied by proofs in the present paper or are standard folklore facts with independent references, and none functions as an unverified load-bearing premise. The only genuinely thin load-bearing point is Lemma 5.4 in characteristic 2, whose proof sketch asserts that the characteristic-zero hypothesis of [GS84] is not needed; this is a proof gap and a correctness risk for Theorem 1.2, but it is not circularity, because Lemma 5.4 is not equivalent to the theorem being proved and is not obtained by fitting to the target result. Overall circularity score: 0.
Assumptions & free parameters
assumptions (4)
- standard math Standard monomials form a k-basis for the Stanley-Reisner ring of a boolean complex, and straightening laws compute products via the ASL relations.
- standard math Reisner-Munkres characterization of Cohen-Macaulayness for boolean complexes via reduced and relative homology vanishing.
- standard math A free Nn-graded module over an Nn-graded connected algebra has an Nn-homogeneous basis.
- domain assumption The invariant ring k[Sd Δ]^{An} of the barycentric subdivision of a simplex is a free k[Θ]-module with basis {1, \tilde D} in characteristic 2.
Cite this review
Pith. "Pith review of Comparing the face rings of a boolean complex and its barycentric subdivision." pith.science (2026). https://pith.science/paper/SOOGMW56
@misc{pith2026250720037,
author = {Pith},
title = {Pith review of: Comparing the face rings of a boolean complex and its barycentric subdivision},
year = {2026},
howpublished = {\url{https://pith.science/paper/SOOGMW56}},
note = {Machine review of arXiv:2507.20037}
}
abstract
We consider the relationship between the Stanley--Reisner ring (a.k.a. face ring) of a simplicial or boolean complex $\Delta$ and that of its barycentric subdivision. These rings share a distinguished parameter subring. S. Murai asked if they are isomorphic, equivariantly with respect to the automorphism group $\operatorname{Aut}(\Delta)$, as modules over this parameter subring. We show that, in general, the answer is no, but for Cohen--Macaulay complexes in characteristic coprime to $|\operatorname{Aut}(\Delta)|$, it is yes, and we give an explicit construction of an isomorphism. To give this construction, we adapt a pair of tools introduced by A. Garsia in 1980. The first one transfers bases from a Stanley--Reisner ring to closely related rings of which it is a Gr\"obner degeneration, and the second identifies bases to transfer.
Figures
Figures from the paper (2 more)
Reviewed August 6, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.