REVIEW 3 major objections 5 minor 14 references
Bounded powers of edge ideals: symmetric exchange binomials
T0 review · 3 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash
Pith's one-line read For every path graph and every bound vector, the toric ideal of the bounded-power polymatroid is generated by symmetric exchange binomials.
desk verdict A genuine extension of the symmetric-exchange program with a promising new induction, but the path theorem's Step 1 rests on an unproved containment that the referee should check carefully. 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 carrying object is the symmetric exchange binomial: when two bases $x^a$ and $x^b$ of a polymatroid allow an exchange of $x_\xi$ for $x_\rho$ such that both $x_\rho x^a/x_\xi$ and $x_\xi x^b/x_\rho$ are bases, the binomial $z_iz_j-z_{i'}z_{j'}$ is a generator candidate for the toric ideal. The new machinery for the path theorem is the $\tau$-invariant, defined as the maximum, over edges of the path and pairs of variables appearing in a monomial, of the difference between the exponents that the two corresponding bases put on that edge. The proof is a double induction: the strong step replaces a monomial with $\tau\ge2$ by one of smaller $\tau$ using a symmetric exchange binomial, and the base step $\tau\le1$ factors the monomial into subpath polymatroids, where a cited product theorem for polymatroids with the strong exchange property gives the required generation. The ordinary-power theorem is carried by the fact that the generators of $I(G)$ form a matroid exactly when $G$ is complete multipartite, together with a quoted theorem on Veronese-type subrings that preserves symmetric-exchange generation under taking products.
What would settle it
Compute $J_{B(P_6,c)}$ for a small path, for instance $P_6$ with $c=(2,1,2,1,1,2)$, in a computer algebra system and compare its minimal binomial generators with the ideal generated by all symmetric exchange binomials of $B(P_6,c)$; a single minimal generator that is not a symmetric exchange binomial would refute Theorem 5.4. The Step 1 containment (7) can also be checked directly on such an example, since the paper leaves that containment unproved.
Extended reading notes
Core claim
The paper's central claim is that for a finite graph $G$ and exponent vector $c$, the polymatroid $B(G,c)$---the set of monomial generators of the highest nonzero bounded power $(I(G)^{\delta_c(I(G))})_c$---has a toric ideal generated by symmetric exchange binomials, and the paper establishes this for the classes listed in the summary. The proof for paths uses a new invariant: for a monomial in the toric ring, let $\tau$ be the largest difference between the exponents that two occurring bases put on the same edge of the path. When $\tau\ge 2$, a symmetric exchange binomial rewrites the monomial to one with smaller $\tau$; when $\tau\le 1$, the monomial factors through products of smaller path polymatroids whose toric ideals are known to be zero, and the product theorem for strong-exchange polymatroids finishes the job. For the other classes, the argument reduces the bounded-power polymatroid to a product of polymatroids with the strong exchange property, whose toric ideals are already known to be symmetric-exchange generated. The paper also shows that if $I(G)^q$ is polymatroidal for any $q\ge1$, then $G$ is complete multipartite, and in that case all powers are polymatroidal with toric ideals generated by symmetric exchange binomials.
Load-bearing premise
The path theorem's Step 1 relies on the containment displayed as (7)---that every bounded-power monomial with $\tau\le1$ factors into monomials on the subpaths separated by the coordinates of $c''$ that exceed one---which is asserted without proof; the classification theorems for the other graph classes similarly rest on strong-exchange classifications in two companion preprints rather than on proofs inside this paper.
Editorial extensions
If this is right
- For every path graph and every nonnegative exponent vector, the toric ideal of the bounded-power polymatroid is generated by symmetric exchange binomials, settling the general conjecture for the whole path family.
- For every graph whose connected components are complete multipartite graphs with a matching deleted, the same generation result holds, so the conjecture is confirmed on a large family beyond paths.
- Whenever the bounded power has degree two, the generation result holds for every graph, so every polymatroid of degree four arising from a bounded edge-ideal power is covered.
- Ordinary powers of an edge ideal are polymatroidal only for complete multipartite graphs, and for those graphs every power has a symmetric-exchange-generated toric ideal.
- If a graph with a leaf has the generation property for every bound, then the graph obtained by deleting that leaf also has the property, giving a downward induction tool for future families.
Reading between the lines
- The $\tau$-invariant induction for paths is likely to transfer to other graphs with a linear or tree-like structure, where removing a vertex separates the graph into smaller pieces; the paper does not make this claim.
- The leaf-deletion theorem only runs from a graph with a leaf to the smaller graph, so an inductive proof for all trees would need the opposite direction, showing that adding a leaf preserves the property; this is not established here.
- The complete multipartite characterization suggests that counterexamples to the general conjecture, if any, cannot come from ordinary powers of non-complete-multipartite graphs, because those powers are not polymatroidal at all.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper studies toric ideals of discrete polymatroids, in particular those arising as the sets of monomial generators of bounded powers of edge ideals. The main results are: Theorem 3.1, classifying when the minimal generators of powers of an edge ideal form a polymatroid; Theorem 4.3, giving a class of graphs for which the bounded-power polymatroid has a toric ideal generated by symmetric exchange binomials via the strong exchange property; Theorem 5.2, treating the case δ_c(I(G)) = 2; Theorem 5.4, the main theorem, asserting that for every path P_n and every capacity vector c, the toric ideal of B(P_n,c) is generated by symmetric exchange binomials; and Theorem 5.5, a leaf-removal reduction. The proof of Theorem 5.4 proceeds by induction on a new invariant τ(u), with a base step at τ = 1 and a reduction step for τ ≥ 2.
Significance. If Theorem 5.4 is correct, it establishes the Herzog-Hibi-White conjecture for all bounded powers of path edge ideals, a natural and previously open family. The τ-invariant induction is a genuinely new technique and could plausibly extend to other graphs. Theorem 5.2 also gives a broad non-trivial class of degree-4 polymatroids. The paper is clearly written and the algebraic context is standard. Its main limitations are that part of the argument in the key τ = 1 base step is asserted rather than proved, and that some classifications on which Theorem 5.2 rests are cited from two companion preprints rather than proved here.
major comments (3)
- [§5, Theorem 5.4, Step 1] The containment (7), w_j ∈ ∏_{i=1}^{h+1} B(G_i,c''_i) ⊂ B(P_n,c''), is load-bearing and is not proved. From τ(u),τ(u') ≤ 1 the authors only deduce that each w_j is a product of distinct edges of P_n. That squarefreeness alone does not imply that the restriction of w_j to each segment G_i is a base of B(G_i,c''_i), i.e. has degree exactly δ_{c''_i}(I(G_i)). Nor is the second inclusion automatic: it requires the degree identity δ_{c''}(I(P_n)) = Σ_i δ_{c''_i}(I(G_i)), which is neither stated nor proved. Without (7), the τ = 1 induction base fails and Theorem 5.4 is not established. The authors should supply a proof of both inclusions, or give a precise counterexample if they fail.
- [§3, Theorem 3.1] The proof of the equivalence in Theorem 3.1 is too compressed and appears incomplete. After choosing x,y,z,w with {x,y} ∉ E(G), {x,z} ∈ E(G), {y,w} ∈ E(G) and {y,z} ∉ E(G), the text asserts that the exchange property forces one of the two listed alternatives, and then declares each impossible. However, the exchange property applied to (xz)^q and (yw)^q only yields a single monomial x_ρ(xz)^q/x_ξ, and its impossibility depends on whether the other factors, such as xw or zw, are edges of G. The proof does not rule out the case where {x,w} or {z,w} is an edge, so the contradiction is not immediate. This theorem is not needed for Theorem 5.4, but it is a stated result and needs a complete proof.
- [§5, Theorem 5.2] The proof of Theorem 5.2 is conditional on the companion preprint [8] for the cases G = C_5, G = K_{2,2,1} - M, and G = K_{2,1,1,1} - M, and the case analysis also contains compressed steps such as 'It then follows easily' in Subcase 1.2. If [8] is not yet available or not accepted, the theorem is not self-contained. The authors should make explicit which statements from [8] are used and, ideally, include the needed cases in this paper or state the dependence clearly in the introduction.
minor comments (5)
- [§4, Corollary 4.2] There is a typo: 'trnasversal' should be 'transversal'.
- [§4, Theorem 4.1] The phrase 'the the toric ideal' contains a duplicate article.
- [§5, Step 1, equations (4) and (6)] The summation index i is used in formulas for r_k and r'_k, making the displayed formulas confusing; the index should be k throughout.
- [§5, Theorem 5.5] In the proof, the homomorphism π' is written K[z_1,...,z_p] → K[v_1,...,v_p], but the final sentence of the setup says '1 ≤ i ≤ q' where q is the number of variables in the larger ring; this should be p.
- [§1, Theorem 1.1] The attribution to [3] is fine, but the sentence 'In the proof of [3, Theorem 5.3 (a)], the following result is essentially proved' could be clarified by stating explicitly whether Theorem 1.1 is proved as a standalone statement in [3] or is an extracted consequence.
Circularity Check
No circularity: Theorem 5.4 is proved by a self-contained tau-induction; the self-citations to companion preprints supply independent external classifications, not the paper's conclusion.
full rationale
The derivation chain contains no step in which the paper's conclusion is fed back as an input. Theorem 5.4 is proved directly by induction on the tau-invariant: Step 2 reduces any binomial with tau at least 2 to binomials of smaller tau using symmetric exchange binomials from the polymatroid structure, and Step 1 handles tau = 1 by reducing to capacities c'' whose components are 0/1, where Lemma 5.3 and Corollary 4.2 give generation by symmetric exchange binomials. None of these steps assumes the target theorem. The principal external dependencies are the authors' companion preprints [7,8], which classify graphs whose bounded-power polymatroids satisfy the strong exchange property, and [3,10], which imply that strong exchange yields generation by symmetric exchange binomials. Those cited results are stronger than or independent of the conjecture being proved here, and they are externally falsifiable classifications rather than restatements of the present conclusion; hence they are real evidence and do not make the reasoning circular. I also flag, as a correctness concern rather than a circularity, that the Step 1 containment (7), w_j in product_i B(G_i,c''_i) subset B(P_n,c''), is asserted without proof, and the equality B(P_n,c'') = {v_j / prod_k e_k^{r_k} : j=1..s} is not demonstrated. If that containment or equality fails, the tau = 1 induction step would collapse; however, that would be a gap in the proof, not a circular reduction, because the theorem's conclusion is not assumed as an input.
Assumptions & free parameters
assumptions (5)
- standard math Published results on discrete polymatroids: K[B] is normal and Cohen-Macaulay; every polymatroid satisfies the symmetric exchange property; strong exchange property implies toric ideal generated by symmetric exchange binomials (Herzog and Hibi, [3], [4]).
- standard math Shibuta's theorem [13, Theorem 3.8]: toric ideals of high Veronese subrings and products are generated by symmetric exchange binomials, used in Theorem 3.1.
- standard math Nicklasson's product theorem [10, Theorem 2.5], used in Theorem 4.1 and Corollary 4.2.
- domain assumption Companion results [7,8] classifying graphs with the strong exchange property for all c (K-M graphs, cycles, trees, unicyclic graphs).
- standard math Graph-theoretic facts: Cameron-Walker graph structure [5] and chordal graph clique-independent decomposition [2, Theorem 2] used in the proof of Theorem 5.2.
Cite this review
Pith. "Pith review of Bounded powers of edge ideals: symmetric exchange binomials." pith.science (2026). https://pith.science/paper/K3SUIH6H
@misc{pith2026250711815,
author = {Pith},
title = {Pith review of: Bounded powers of edge ideals: symmetric exchange binomials},
year = {2026},
howpublished = {\url{https://pith.science/paper/K3SUIH6H}},
note = {Machine review of arXiv:2507.11815}
}
read the original abstract
It has been conjectured that the toric ideal of the base ring of a discrete polymatroid is generated by symmetric exchange binomials. In the present paper, we give several classes of discrete polymatroids which yield toric ideals generated by symmetric exchange binomials. Especially, we are interested in the discrete polymatroids arising from bounded powers of edge ideals of finite graphs.
Reference graph
Works this paper leans on
-
[3]
J. Herzog and T. Hibi, Discrete polymatroids, J. Algebraic Combin. 16 (2002), 239–268
work page 2002
-
[8]
Bounded powers of edge ideals: The strong exchange property
T. Hibi and S. A. Seyed Fakhari, Bounded powers of edge ideals: The strong exchange prop- erty, arXiv:2506.03480
-
[1]
Blum, Base-sortable matroids and Koszulness of semigroup rings, Europ
S. Blum, Base-sortable matroids and Koszulness of semigroup rings, Europ. J. Combina. 22 (2001), 937–951
work page 2001
-
[2]
Cameron, Induced matchings, Discrete Appl
K. Cameron, Induced matchings, Discrete Appl. Math. 24 (1989), 97–109
work page 1989
-
[4]
Monomial Ideals,
J. Herzog and T. Hibi, “Monomial Ideals,” GTM 260, Springer, 2011
2011
-
[5]
T. Hibi, A. Higashitani, K. Kimura and A. B. O’Keefe, Algebraic study on Cameron–Walker graphs, J. Algebra 422 (2015), 257–269
work page 2015
-
[6]
Bounded powers of edge ideals: regularity and linear quotients
T. Hibi and S. A. Seyed Fakhari, Bounded powers of edge ideals: regularity and linear quo- tients, Proc. Amer. Math. Soc. , to appear, arXiv:2502.01768
-
[7]
Bounded powers of edge ideals: Gorenstein toric rings
T. Hibi and S. A. Seyed Fakhari, Bounded powers of edge ideals: Gorenstein toric rings, arXiv:2504.21760
Show all 14 references
-
[9]
Kashiwabara, The toric ideal of a matroid of rank 3 is generated by quadrics, Electr
K. Kashiwabara, The toric ideal of a matroid of rank 3 is generated by quadrics, Electr. J. Combina. 17 (2010), #28
2010
-
[10]
Nicklasson, Products of polymatroids with the strong exchange property, J
L. Nicklasson, Products of polymatroids with the strong exchange property, J. Pure and Appl. Alg. 226 (2022), #106967
2022
-
[11]
Ohsugi and T
H. Ohsugi and T. Hibi, Compressed polytopes, initial ideals and complete multipartite graphs, Illinois J. Math. 44 (2000), 391–406
2000
-
[12]
Schweig, Toric ideals of lattice path matroids and polymatroids, J
J. Schweig, Toric ideals of lattice path matroids and polymatroids, J. Pure and Appl. Algebra 215 (2011), 2660–2665
2011
-
[13]
Shibuta, Toric ideals for high Veronese subrings of toric algebras, Illinois J
T. Shibuta, Toric ideals for high Veronese subrings of toric algebras, Illinois J. Math. 55 (2011), 895–905
2011
-
[14]
White, A unique exchange property for bases, Linear Algebra Appl
N. White, A unique exchange property for bases, Linear Algebra Appl. 31 (1980), 81–91. (Takayuki Hibi) Department of Pure and Applied Mathematics, Graduate School of Information Science and Technology, Osaka University, Suita, Osaka 565–0871, Japan Email address : hibi@math.sc...
1980
Reviewed August 6, 2026 · model on record in the stance chip above.
Discussion (0). Sign in to comment.