REVIEW 1 major objections 4 minor 15 references
Edge rings of bipartite graphs with linear resolutions
T0 review · 1 major / 4 minor · reviewed 2026-08-14 · deepseek-v4-flash
Pith's one-line read A finite connected simple bipartite graph whose edge ring has a q-linear resolution, q ≥ 3, must be a hypersurface.
desk verdict Proves the bipartite case of the q-linear resolution conjecture; solid result with a terse classification step that needs expanding. 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 mechanism is the pair of inequalities $\deg(P_{G'}) \leq \deg(P_G)$ for subgraphs and $\mathrm{reg}(K[G]) \geq \deg(P_G)$, together with the identity $\deg(P) = \dim P + 1 - \operatorname{codeg}(P)$, where $\operatorname{codeg}(P)$ is the smallest r with an interior lattice point in $rP$. To get degree at least q for each forbidden configuration, the author writes an explicit convex combination of edge vectors with coefficients 1/3 and 2/3 that lands in the interior of $rP_G$ for suitable r; this certifies $\operatorname{codeg}(P_G) \leq r$, hence $\deg(P_G) \geq \dim P_G + 1 - r \geq q$. The convex combinations are the load-bearing calculations, and the two families $G^{(e)}_{k,m}$ and $G^{(o)}_{k,m}$ are the parameterized subgraphs needed to cover the overlapping-cycle case.
What would settle it
Build or search for a connected bipartite graph with two distinct 2q-cycles, no even cycle shorter than 2q, and whose union is not isomorphic to any $G^{(e)}_{k,m}$ or $G^{(o)}_{k,m}$; then compute the degree of the h*-polynomial of its edge polytope. A degree below q would refute the theorem, while a degree at least q in such a graph would show only that the classification step in the proof needs repair, not that the theorem fails.
Extended reading notes
Core claim
The central theorem, Theorem 0.2, states that for a finite connected simple bipartite graph G and a field K, if the edge ring K[G] has a q-linear resolution with q ≥ 3, then K[G] is a hypersurface. Equivalently, the toric ideal I_G is principal and G has exactly one even cycle, of length 2q; the single binomial generating I_G is the binomial f_C attached to that cycle. The proof runs by contradiction: a q-linear resolution implies I_G is generated in degree q, hence by Lemma 1.2 G has no even cycle shorter than 2q and every generator comes from a 2q-cycle. Assuming two distinct 2q-cycles exist, the paper builds a subgraph (disjoint cycles, cycles sharing one vertex, or one of the two families $G^{(e)}_{k,m}$ and $G^{(o)}_{k,m}$) and exhibits an interior lattice point in a small dilation of its edge polytope; the codegree identity then gives $\deg(P_G) \geq q$. Since regularity is at least this degree, the resolution cannot be q-linear, a contradiction.
Load-bearing premise
The load-bearing premise is that every connected bipartite subgraph made of a 2q-cycle plus a path joining two of its vertices, with no odd cycle and no even cycle shorter than 2q, is one of the two explicitly drawn families $G^{(e)}_{k,m}$ or $G^{(o)}_{k,m}$; the paper asserts this without giving the classification argument.
Editorial extensions
If this is right
- For q ≥ 3, any bipartite edge ring with a q-linear resolution has a principal toric ideal; in particular it is a hypersurface and its minimal free resolution has length one.
- Conversely, by the generation statement of Lemma 1.1, a connected bipartite graph with exactly one even cycle of length 2q has edge ring with a q-linear resolution, so the theorem gives a complete combinatorial characterization: q-linear resolutions of bipartite edge rings are exactly the unicyclic graphs whose unique cycle has length 2q.
- The h*-polynomial degree of the edge polytope of any graph with two 2q-cycles and no shorter even cycle is at least q, so regularity is at least q; this gives an obstruction that can be checked from the graph alone.
- The result settles the q ≥ 3 bipartite case of Conjecture 0.1, leaving the non-bipartite case open.
Reading between the lines
- The same codegree-certificate strategy would likely prove the full conjecture for any class of graphs whose toric ideal is generated by cycle binomials; for non-bipartite graphs, one would need analogues of Lemmas 2.3 and 2.4 involving odd-cycle binomials.
- The classification statement that every allowed $C_1 \cup P$ is one of the two drawn families is the only non-explicit step; making it explicit would yield a self-contained proof and might simplify the two parameter families.
- A direct combinatorial reformulation emerges: for bipartite graphs, 'q-linear edge ring' is equivalent to 'exactly one cycle, of length 2q', giving an easy way to identify such edge rings from the graph alone without computing a resolution.
Signed reviews
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper proves Theorem 0.2: for a finite connected simple bipartite graph G whose edge ring K[G] has a q-linear resolution with q >= 3, K[G] is a hypersurface. The proof uses the fact that a q-linear resolution forces I_G to be generated in degree q, which by Lemma 1.2 means G has no even cycles shorter than 2q and I_G is generated by the binomials of its 2q-cycles. The author then shows that if G had two 2q-cycles, a subgraph argument would give deg(P_G) >= q, contradicting reg(K[G]) <= q-1 via Lemma 1.4. The remaining case is that G has exactly one 2q-cycle, and then Lemma 1.2 makes I_G principal, so K[G] is a hypersurface.
Significance. If the proof is completed, this is a clean positive solution of Conjecture 0.1 in the bipartite case, with an elementary combinatorial proof. The main strength is that the numerical certificates in Lemmas 2.1, 2.2, 2.3 and 2.4 are explicit convex combinations of edge vectors, so the degree lower bounds are directly verifiable. The reduction to a unique 2q-cycle is elegant, and the use of the monotonicity deg(P_G') <= deg(P_G) is appropriate. The only substantive weakness is that one load-bearing classification step in the proof of Theorem 0.2 is asserted without proof.
major comments (1)
- [Section 2, proof of Theorem 0.2] The sentence 'Hence G′ is G^(e)_{k,m} or G^(o)_{k,m} which appear in Lemmas 2.3 and 2.4' is asserted without proof, and this assertion is load-bearing: it transfers the degree lower bounds of Lemmas 2.3 and 2.4 from the two explicitly drawn families to an arbitrary pair of overlapping 2q-cycles. The author should supply the missing argument. Concretely, if P is the path whose endpoints w1,w2 lie on C1 and whose internal vertices lie outside C1, let r be the length of the shorter C1-arc between w1 and w2 and let L be the length of P. Bipartiteness forces L and r to have the same parity, and the condition that G′ has no even cycle shorter than 2q gives L+r >= 2q and L+(2q-r) >= 2q. These inequalities imply that (r,L) is either (2k,2m) with k+m >= q, or (2k-1,2m-1) with k+m-1 >= q, which are exactly the parameter ranges used in Lemmas 2.3 and 2.4, and the pair (r,L) determines the isomorphism type of G′. Until this parity/arc-length argument is written out, the main contradiction deg(P_G) >= q is not fully justified.
minor comments (4)
- [Section 2, proof of Theorem 0.2] In the final paragraph, 'no even cycles of length < 2n' should read 'no even cycles of length < 2q'; the symbol n is not defined at that point.
- [Lemma 2.1] The subgraph G′ used in Lemma 2.1 is the disconnected disjoint union of two cycles, while the cited monotonicity lemma is stated for a subgraph of a connected graph. If Lemma 1.3 is only proved for connected subgraphs, the author should either state the version for disconnected subgraphs explicitly or replace G′ by a connected spanning subgraph of G containing the two disjoint cycles, since connectedness of G supplies edges between the cycles.
- [Lemma 2.2] In the proof of Lemma 2.2, 'It follows that dim PGk = 4q − 3' contains a typo; it should be dim P_{G′} = 4q − 3.
- [Lemma 2.3] In the proof of Lemma 2.3, the inequality 'deg(P_{G^(e)_{k,m}}) >= 3q/2 - 2' is stated, but for q = 3 the text writes 'deg >= 3 > 5/2'. Since degrees are integers, the conclusion deg >= 3 is clear, but the comparison to 5/2 is unnecessarily indirect.
Circularity Check
No significant circularity: the proof derives the hypersurface conclusion from standard regularity facts and explicit degree computations, with no fitted input or self-referential reduction.
full rationale
The derivation chain is self-contained for the purposes of circularity analysis. A q-linear resolution implies reg = q-1 and generation in degree q; Lemma 1.2 then gives the absence of even cycles shorter than 2q, and Lemma 1.4 gives deg(P_G) <= q-1. The contradiction argument proves that any second 2q-cycle forces deg(P_G) >= q via explicit Ehrhart interior-point computations in Lemmas 2.1-2.4, so the conclusion that G has exactly one 2q-cycle, hence that I_G is principal and K[G] is a hypersurface, is not assumed as an input. The paper cites Lemma 1.3 from the authors' earlier work, but that lemma is a monotonicity result originally attributed to Stanley and is not an equivalent restatement of the target theorem; Lemma 1.4 comes from independent work of Hofscheier-Katthän-Nill. The only notable weakness is the unproved assertion that the subgraph C1 union P must be one of the two families G^(e) or G^(o), which is an omitted combinatorial classification argument rather than a circular step, since the classification does not presuppose the hypersurface conclusion. No fitted parameter, normalization, or self-defined quantity is renamed as a prediction.
Assumptions & free parameters
assumptions (5)
- standard math Toric ideal of a connected bipartite graph is generated by even-cycle binomials f_C (Lemma 1.1, cited from [5, Cor. 5.12]).
- standard math deg(P_{G'}) ≤ deg(P_G) for every subgraph G' of a connected simple graph G (Lemma 1.3, cited from [7, Cor. 3.2]).
- standard math reg(K[P_G]) ≥ deg(P_G) for a connected simple graph G (Lemma 1.4, cited from [8]).
- standard math Ehrhart theory: codeg(P) = min{r : int(rP) ∩ Z^N nonempty} and deg(P) = dim(P)+1-codeg(P).
- standard math If S/I has a q-linear resolution then I is generated in degree q and reg(S/I)=q-1.
Cite this review
Pith. "Pith review of Edge rings of bipartite graphs with linear resolutions." pith.science (2026). https://pith.science/paper/XDI2XHNI
@misc{pith2026190805678,
author = {Pith},
title = {Pith review of: Edge rings of bipartite graphs with linear resolutions},
year = {2026},
howpublished = {\url{https://pith.science/paper/XDI2XHNI}},
note = {Machine review of arXiv:1908.05678}
}
abstract
Ohsugi and Hibi characterized the edge ring of a finite connected simple graph with a $2$-linear resolution. On the other hand, Hibi, Matsuda and the author conjectured that the edge ring of a finite connected simple graph with a $q$-linear resolution, where $q \geq 3$, is a hypersurface and proved the case $q=3$. In the present paper, we solve this conjecture for the case of finite connected simple bipartite graphs.
Reference graph
Works this paper leans on
-
[1]
J. Biermann, A. O’Keefe and A. V an Tuyl, Bounds on the regu larity of toric ideals of graphs, Adv. in Appl. Math. 85 (2017), 84–102
work page 2017
-
[2]
W . Bruns and J. Herzog, Cohen-Macaulay Rings, Revised Ed ., Cambridge Stud. Adv. Math., vol. 39, Cambridge University Press, Cambridge, 1998
work page 1998
-
[3]
Green-Lazarsfeld Condition for Toric Edge Ideals of Bipartite Graphs
Z. Greif and J. McCullough, Green-Lazarsfeld Condition for Toric Edge Ideals of Bipartite Graphs, arXiv:1908.02744. 7
work page Pith review arXiv 1908
-
[4]
H. T. H´ a, S. Kara and A. O’Keefe. Algebraic properties of toric rings of graphs, Comm. Algebra 47 (2019), 1–16
work page 2019
- [5]
-
[6]
Algebraic Combinatorics on Convex Polytopes,
T. Hibi, “Algebraic Combinatorics on Convex Polytopes, ” Carslaw Publications, Glebe, N.S.W ., Aus- tralia, 1992
work page 1992
-
[7]
T. Hibi, K. Matsuda and A. Tsuchiya, Edge rings with 3-lin ear resolutions, Proc. Amer . Math. Soc. 147 (2019), 3225–3232
work page 2019
-
[8]
J. Hofscheier, L. Katth¨ an and B. Nill, Ehrhart theory of spanning lattice polytopes, Int. Math. Res. Not. IMRN 2018 (2018), 5947–5973
work page 2018
Show all 15 references
-
[9]
K´ alm´ an and A
T. K´ alm´ an and A. Postnikov, Root polytopes, Tutte polynomials, and a duality theorem for bipartite graphs, Proc. Lond. Math. Soc. 114 (2017), 561–588
2017
-
[10]
Ohsugi and T
H. Ohsugi and T. Hibi, Normal polytopes arising from fini te graphs, J. Algebra 207 (1998), 409–426
1998
-
[11]
Ohsugi and T
H. Ohsugi and T. Hibi, Koszul bipartite graphs, Adv. Applied Math. 22 (1999), 25–28
1999
-
[12]
Ohsugi and T
H. Ohsugi and T. Hibi, Toric ideals generated by quadrat ic binomials, J. Algebra 218 (1999), 509–527
1999
-
[13]
R. P . Stanley, A monotonicity property of h-vectors and h∗-vectors, European J. Combin. 14 (1993), 251–258
1993
-
[14]
Gr¨ obner bases and convex polytopes
B. Sturmfels, “Gr¨ obner bases and convex polytopes”, A mer. Math. Soc., Providence, RI, 1996
1996
-
[15]
C. E. V alencia and R. H. Villarreal, Explicit represent ations of the edge cone of a graph, Int. J. Con- temp. Math. Sci. 1 (2006), 53–66. (Akiyoshi Tsuchiya) G RADUATE SCHOOL OF MATHEMATICAL SCIENCES , U NIVERSITY OF TOKYO , KOMABA , M EGURO -KU, T OKYO 153-8914, J APAN E-ma...
2006
Reviewed August 14, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.