Pith. sign in

REVIEW 3 minor 21 references

Nef-partitions arising from unimodular configurations

T0 review · 0 major / 3 minor · reviewed 2026-08-14 · deepseek-v4-flash

Pith's one-line read The paper proves that every spanning unimodular configuration yields a Gorenstein Cayley sum, a reflexive Minkowski sum, and a nef-partition.

desk verdict Clean Gröbner-basis proof of a new family of nef-partitions from unimodular configurations; the only real gap is a true but unproved parity assertion in Lemma 3.5. read the letter →

arxiv 1908.01369 v2 pith:WLK7VUDQ submitted 2019-08-04 math.CO math.AC

classification math.COmath.AC MSC 05A1505C3113P1052B1252B20
keywords reflexivepolytopeGorensteinnef-partitionintegerdecompositionpropertyGröbnerbasisunimodularconfigurationedgetoricideal
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper establishes a general construction of nef-partitions—decompositions of a reflexive polytope into Minkowski summands that contain the origin—from any unimodular configuration A whose polytope PA has exactly the columns of A as lattice points and is spanning. The main theorem says that PA ∗ (−PA) is Gorenstein of index 2 with a regular unimodular triangulation, that PA + (−PA) is reflexive with a regular unimodular triangulation, and that (PA − a) + (−PA + a) is a nef-partition for every lattice point a of PA. The proof is algebraic and explicit: it constructs reduced Gröbner bases with squarefree initial ideals for the toric ideals of the centrally symmetric configuration and of the two Cayley sums, then transfers these triangulations to the Minkowski sums. Because nef-partitions are the combinatorial data behind mirror pairs of Calabi–Yau complete intersections, the theorem yields large new families of explicit mirrors. For graph edge polytopes, the hypotheses hold whenever all odd cycles share a common vertex, and the zero-augmented version works exactly for bipartite graphs.

What carries the argument

The load-bearing object is the toric ideal IA± of the centrally symmetric configuration A± = (a1,...,an, −a1,...,−an, 0), studied through reverse lexicographic Gröbner bases. Theorem 1.2 shows its reduced Gröbner basis is {xiyi − z²} ∪ {g1,...,gs}, where each gi is a squarefree binomial and no initial monomial involves x1 or y1. The same Gröbner basis, with z² replaced by x1y1, becomes the reduced Gröbner basis of the toric ideal of the Cayley sum PA ∗ (−PA) (Theorem 1.3), and with z² replaced by x0y0 it does the same for PA0 ∗ (−PA0) (Theorem 1.4). Squarefree initial ideals give regular unimodular triangulations of the Cayley sums, and Lemma 2.2 transfers this property to the Minkowski sum when the summands are spanning. Hilbert-function comparison then identifies h*(PA ∗ (−PA), t) with h(K[A±], t)/(1 + t); the known palindromicity of h(K[A±], t) forces h* to be palindromic of degree d − 1, which is exactly the criterion for Gorenstein index 2.

What would settle it

Compute the h*-polynomial of PA ∗ (−PA) for a small spanning unimodular configuration satisfying PA ∩ Z^d = {columns}, for instance the edge polytope of a graph whose odd cycles all share a vertex, and check whether it is palindromic of degree d − 1; if any such computation gives a non-palindromic h*, or if PA + (−PA) fails to be reflexive, the theorem is false. A direct computer search over all spanning unimodular configurations in small dimensions would settle the universality claim.

Watch

Extended reading notes

Core claim

Let A = (a1,...,an) be a unimodular configuration: all nonzero maximal minors have the same absolute value. Under the hypotheses PA ∩ Z^d = {a1,...,an} and PA spanning, Theorem 0.1 proves three statements: (1) the Cayley sum PA ∗ (−PA) is Gorenstein of index 2 and admits a regular unimodular triangulation; (2) the Minkowski sum PA + (−PA) is reflexive, admits a regular unimodular triangulation, and consequently (PA − a) + (−PA + a) is a nef-partition for every a in PA ∩ Z^d; and (3) the lattice-point equality (PA ∩ Z^d) + (−PA ∩ Z^d) = (PA + (−PA)) ∩ Z^d holds. Theorem 0.2 is the analogous statement for A0 = (A, 0), with the conclusion that PA0 + (−PA0) itself is a nef-partition. The central algebraic fact behind these results is that the toric ideal of the centrally symmetric configuration A± has a squarefree initial ideal with respect to a reverse lexicographic order, and that the reduced Gröbner basis has the explicit form {xiyi − z²} ∪ {g1,...,gs}; the same binomials, with z² replaced by x1y1 or x0y0, give squarefree initial ideals for the Cayley sums.

Load-bearing premise

The whole construction depends on the polytope being spanning—its lattice points must generate the ambient lattice—and without that, the squarefree initial ideal of the Cayley sum cannot be guaranteed to transfer into a unimodular triangulation of the Minkowski sum; the graph application shows this is a real boundary, since the zero-augmented edge polytope is spanning exactly for bipartite graphs.

Editorial extensions

If this is right

  • Every unimodular configuration satisfying the two hypotheses produces a reflexive polytope PA + (−PA) with a regular unimodular triangulation, hence an IDP polytope whose Ehrhart h*-polynomial is palindromic.
  • Every such configuration produces a nef-partition (PA − a) + (−PA + a) for each lattice point a, giving explicit input for mirror-pair constructions of Calabi–Yau complete intersections.
  • For any graph whose odd cycles pairwise share a vertex, the edge polytope satisfies the hypotheses, so PAG + (−PAG) is reflexive and (PAG − a) + (−PAG + a) is a nef-partition for every lattice point a of the edge polytope.
  • For any connected bipartite graph, the zero-augmented edge polytope P(AG)0 satisfies the hypotheses, so P(AG)0 + (−P(AG)0) is a reflexive nef-partition and the h*-polynomial identity h*(P(AG)0 ∗ (−P(AG)0), t) = (1 + t)h*(PAG ∗ (−PAG), t) holds.
  • The equality (PA ∩ Z^d) + (−PA ∩ Z^d) = (PA + (−PA)) ∩ Z^d is a concrete instance of the integer decomposition property, directly relevant to the question of when Minkowski sums of lattice polytopes decompose integer points.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The proof uses only the existence of a squarefree initial ideal for the centrally symmetric configuration together with the spanning condition, so configurations that are not unimodular but whose A± still has a squarefree initial ideal may admit the same conclusions; unimodularity may be sufficient rather than necessary.
  • For bipartite graphs, the factorization h*(PA0 ∗ (−PA0), t) = (1 + t)h*(PA ∗ (−PA), t) suggests a general combinatorial explanation: adding the origin as a lattice point may split the Cayley-sum lattice points into two copies, giving a direct bijective proof of Corollary 2.3.
  • One could compute a full regular unimodular triangulation of PA + (−PA) directly from the explicit Gröbner basis in Theorem 1.2, yielding an algorithmic way to list all maximal simplices and hence the entire Ehrhart series for these polytopes.
  • For non-bipartite graphs, P(AG)0 fails to be spanning; testing whether PA0 + (−PA0) can still be reflexive in some non-spanning cases would locate exactly where the spanning assumption is needed.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

0 major / 3 minor

Summary. The paper studies reflexive and Gorenstein polytopes constructed from unimodular configurations. The main results, Theorems 0.1 and 0.2, assert that for a unimodular configuration A whose polytope PA has exactly the columns as lattice points and is spanning, the Cayley sum PA ∗ (−PA) is Gorenstein of index 2 with a regular unimodular triangulation, while the Minkowski sum PA + (−PA) is reflexive with a regular unimodular triangulation and gives a nef-partition; an analogous statement is proved for A0 = (A, 0). The proofs proceed by explicitly constructing reduced Gröbner bases for the toric ideals of A±, PA ∗ (−PA), and PA0 ∗ (−PA0) in Theorems 1.2–1.4, then comparing h-polynomials of the associated initial ideals to obtain the Gorenstein and reflexivity conclusions. Section 3 applies the main theorems to edge polytopes of finite graphs, characterizing when the configuration with the zero column is spanning and thereby obtaining nef-partitions for graphs whose odd cycles pairwise meet and for bipartite graphs.

Significance. If the results are correct, the paper provides a large and explicit family of nef-partitions and reflexive polytopes arising from unimodular configurations, with regular unimodular triangulations and the integer decomposition property as byproducts. The explicit Gröbner bases are a concrete strength: they give algorithmic control over the triangulations and make the IDP and h*-polynomial comparisons transparent. The spanning hypothesis is also carefully delineated, and the graph-theoretic characterization in Section 3 shows that this hypothesis is natural and often sharp. Conditional on the two cited external ingredients, Proposition 1.1 from [17] and Lemma 2.2 from [13], the central derivation is clean and complete. I found no load-bearing error in Theorems 0.1–0.2 or in the Gröbner-basis arguments of Section 1.

minor comments (3)
  1. [3, Lemma 3.5] The assertion that every lattice point of P_{(AG)0} has even coordinate sum when G is non-bipartite is stated without proof. This is true: each edge vector has coordinate sum 2, the origin has coordinate sum 0, and every convex combination therefore has coordinate sum 2λ, which is an even integer when the point is integral. Please add this one-sentence justification.
  2. [2, proofs of Theorems 0.1 and 0.2] The dimension statement that PA ∗ (−PA) has dimension d may be initially confusing because the Cayley sum is embedded in R^{d+1}. The point is that the columns of a configuration lie in an affine hyperplane not passing through the origin, so PA has affine dimension d−1 under the rank-d assumption. Stating this explicitly before the dimension count would improve readability.
  3. [1, Theorem 1.2] The proof of Theorem 1.2 is dense, especially the final argument using circuits from [11, Lemma 4.32 and Theorem 4.35]. A short indication of why the divisibility contradiction at the end forces q to be squarefree would help the reader follow the reduced-Gröbner-basis argument.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the central derivation is a genuine reduction to prior parameter-free theorems, not to its own conclusion.

full rationale

The derivation chain in Theorems 0.1 and 0.2 is not circular. The authors prove that the Cayley-sum toric ideals I_{PA*(-PA)} and I_{PA0*(-PA0)} have squarefree initial ideals (Theorems 1.3 and 1.4) by analyzing the reduced Grobner basis of I_{A±}; this analysis uses Proposition 1.1 from the authors' earlier paper [17], but that proposition is a general, parameter-free statement about centrally symmetric configurations of every unimodular matrix, and its hypotheses do not include Gorensteinness, reflexivity, or nef-partitions of PA ± (-PA). The later transfer from Cayley sums to Minkowski sums is made by Lemma 2.2 from [13], again a general Cayley-sum theorem about spanning polytopes, not a restatement of the present conclusions. The h*-polynomial comparisons then genuinely identify Gorenstein index 2 and reflexivity, with the squarefree initial ideals supplying regular unimodular triangulations and IDP; no fitted parameter is renamed as a prediction, and no quantity is defined in terms of the quantity it is used to prove. In Section 3, the graph applications invoke independent classifications and a short spanning argument. The only notable gap is the unproved parity assertion in Lemma 3.5, which is an omission in exposition rather than a circularity; the assertion is true and is not load-bearing for the main theorems. Consequently no circular step is exhibited.

Assumptions & free parameters 0 free parameters · 6 assumptions · 0 invented entities

The paper introduces no free parameters and no new entities. The theorems are universal over the class of unimodular configurations. The derivation rests on several prior theorems, two of which are from the same research group; all are parameter-free and do not assume the target result. One auxiliary lattice-point fact in Lemma 3.5 is used without proof.

assumptions (6)
  • standard math Proposition 1.1: for a unimodular matrix A, the initial ideal of I_{A±} is squarefree with respect to a reverse lexicographic order with z smallest, and K[A±] is normal and Gorenstein with palindromic h-polynomial of degree d.
    Cited from [17] and used throughout Section 1 and Section 2 as the foundation for the Grobner basis computations and h-polynomial comparisons.
  • standard math Lemma 2.2 [13, Theorem 2.2]: if P1,...,Pr are spanning lattice polytopes and the toric ideal of their Cayley sum has a squarefree initial ideal, then the toric ideal of the Minkowski sum has a squarefree initial ideal, both have regular unimodular triangulations, and both are IDP.
    This is the bridge from the Grobner basis results to IDP and to the equality h* = h for the Cayley sum and Minkowski sum; it appears in the proofs of Theorems 0.1 and 0.2.
  • standard math A lattice polytope is Gorenstein of index r if and only if its h*-polynomial has degree dim(P) + 1 - r and is palindromic.
    Cited from [8] and used in Section 2 to convert the palindromic h*-polynomial of PA * (-PA) into the Gorenstein index 2 conclusion.
  • standard math Batyrev-Nill Theorem 2.6: P1 + ... + Pr is Gorenstein of index 1 if and only if the Cayley sum P1 * ... * Pr is Gorenstein of index r.
    Used in Section 2 to pass from the Gorenstein index 2 property of the Cayley sum to reflexivity of PA + (-PA) and PA0 + (-PA0).
  • standard math Lemma 3.2: the edge polytope of a finite connected graph is spanning.
    Cited from [12] and used in Section 3 to apply Theorem 0.1 to edge polytopes of graphs whose odd cycles pairwise intersect.
  • ad hoc to paper For a non-bipartite connected graph G, every lattice point of P_{(AG)0} has even coordinate sum.
    Asserted without proof in Lemma 3.5 to show that P_{(AG)0} is not spanning for non-bipartite graphs. The claim is true, but the paper omits the argument that integer points in this 0/1 polytope are the origin and the edge vectors.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Nef-partitions arising from unimodular configurations." pith.science (2026). https://pith.science/paper/WLK7VUDQ

@misc{pith2026190801369,
  author       = {Pith},
  title        = {Pith review of: Nef-partitions arising from unimodular configurations},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/WLK7VUDQ}},
  note         = {Machine review of arXiv:1908.01369}
}
abstract

Reflexive polytopes have been studied from viewpoints of combinatorics, commutative algebra and algebraic geometry. A nef-partition of a reflexive polytope $\mathcal{P}$ is a decomposition $\mathcal{P}=\mathcal{P}_1+\cdots+\mathcal{P}_r$ such that each $\mathcal{P}_i$ is a lattice polytope containing the origin. Batyrev and van Straten gave a combinatorial method for explicit constructions of mirror pairs of Calabi-Yau complete intersections obtained from nef-partitions. In the present paper, by means of Gr\"{o}bner basis techniques, we give a large family of nef-partitions arising from unimodular configurations.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

21 extracted references · 19 canonical work pages

  1. [17]

    Ohsugi and T

    H. Ohsugi and T. Hibi. Centrally Symmetric Configuratio ns of Integer Matrices. Nagoya Math. J. , 216:153–170, 2014

  2. [13]

    T. Hibi, H. Ohsugi, and A. Tsuchiya. Integer decomposit ion property for Cayley sums of order and stable set polytopes. Michigan Math. J. to appear

  3. [1]

    V . V . Batyrev. Dual polyhedra and mirror symmetry for Calabi-Yau hypersurfaces in toric varieties. J. Algebraic Geom., 3:493–535, 1994

  4. [2]

    V . V . Batyrev and D. Juny. Classification of Gorenstein toric del Pezzo varieties in arbitrary dimension. Mosc. Math. J., 10:285–316, 2010

  5. [3]

    V . V . Batyrev and B. Nill. Combinatorial aspects of mirro r symmetry. Contemp. Math. , 452:35–66, 2008

  6. [4]

    V . V . Batyrev and D. van Straten. Generalized hypergeometric functions and rational curves on Calabi- Yau complete intersections in toric varieties. Comm. Math. Phys. , 168(3):493–533, 1995

  7. [5]

    L. A. Borisov. Towards the Mirror Symmetry for Calabi-Ya u Complete Intersections in Gorenstein Toric Fano Varieties. math.AG/0308216, 1993

  8. [6]

    Bruns and J

    W . Bruns and J. Gubeladze. Polytopes, rings, and K-theory . Springer Monographs in Mathematics. Springer, Dordrecht, 2009

Show all 21 references
  1. [7]

    D. A. Cox, J. B. Little, and H. K. Schenck. T oric varieties, volume 124 of Graduate Studies in Math- ematics. American Mathematical Society, Providence, RI, 2011

  2. [8]

    De Negri and T

    E. De Negri and T. Hibi. Gorenstein algebras of Veronese t ype. J. Algebra, 193:629–639, 1997

  3. [9]

    E. Ehrhart. Sur les poly` edres rationnels homoth´ etiqu es ` a n dimensions. C. R. Acad. Sci. Paris , 254:616–618, 1962

  4. [10]

    Herzog and T

    H. Herzog and T. Hibi. Monomial Ideals. Graduate Text in Mathematics. Springer, 2011

  5. [11]

    Herzog, H

    T. Herzog, H. Hibi and H. Ohsugi. Binomial Ideals. Graduate Text in Mathematics. Springer, 2018

  6. [12]

    T. Hibi, K. Matsuda, and A. Tsuchiya. Edge rings with 3-l inear resolutions. Proc. Amer . Math. Soc., 147(8):3225–3232, 2019

  7. [14]

    Joswig and K

    M. Joswig and K. Kulas. Tropical and ordinary convexity combined. Adv. Geom., 10:333–352, 2010

  8. [15]

    T. Oda. Problems on Minkowski sums of convex lattice pol ytopes. arXiv:0812.1418, 2008

  9. [16]

    Ohsugi and T

    H. Ohsugi and T. Hibi. Normal polytopes arising from fini te graphs. J. Algebra, 207:409–426, 1998

  10. [18]

    Schrijver

    A. Schrijver. Theory of Linear and Integer Programing . John Wiley & Sons, 1986

  11. [19]

    R. P . Stanley. Decompositions of rational convex polyt opes. Annals of Discrete Math. , 6:333–342, 1980

  12. [20]

    Sturmfels

    B. Sturmfels. Gr¨obner Bases and Convex Polytopes, volume 8 of University Lecture Series. American Mathematical Society, Providence, RI, 1996

  13. [21]

    Tsuchiya

    A. Tsuchiya. Cayley sums and Minkowski sums of 2-convex -normal lattice polytopes. arXiv:1804.10538, 2018. 11 HIDEFUMI OHSUGI , D EPARTMENT OF MATHEMATICAL SCIENCES , S CHOOL OF SCIENCE AND TECH - NOLOGY , K WANSEI GAKUIN UNIVERSITY , S ANDA , H YOGO 669-1337, J APAN E-mail ad...

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.