Preserving Hodge Vectors of Lattice Polytopes

Benjamin Nill; Vadym Kurylenko

REVIEW 3 major objections 5 minor 4 references

Preserving Hodge Vectors of Lattice Polytopes

T0 review · 3 major / 5 minor · reviewed 2026-08-02 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read The Hodge vector of a Cayley polytope equals the mixed volume of its first k factors times the Hodge vector of the projection of the remaining factor.

desk verdict A mostly solid new mechanism for preserving Hodge vectors under Cayley constructions; the core formula checks out, but two deferred proofs—one load-bearing—should be addressed before publication. read the letter →

arxiv 2602.20765 v2 pith:DO7KQPAS submitted 2026-02-24 math.CO math.AG

Vadym Kurylenko , Benjamin Nill This is my paper · ORCID

classification math.COmath.AG MSC 52B2014M25

keywords Hodgevectorlocalh*-polynomialCayleypolytopemixedvolumeLawrencetwistthinHodge-Delignepolynomialbivariate

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 formula for the Hodge vector — the local h*-vector stripped of leading and trailing zeroes — of a Cayley polytope built from lattice polytopes P1,...,Pk,Q. When P1,...,Pk lie in a k-dimensional subspace U, the Hodge vector of P1*...*Pk*Q equals the mixed volume of P1,...,Pk times the Hodge vector of the projection of Q along U. This yields a systematic way to produce infinitely many high-dimensional lattice polytopes with the same Hodge vector that are not free joins, a phenomenon impossible in classical Ehrhart theory. The proof uses a closed formula for the Hodge-Deligne polynomial of a complete intersection in the torus, expressed through bivariate h*-polynomials. The construction also generates many new thin polytopes and explains the thinness of B_k-polytopes.

What carries the argument

The central mechanism is the bivariate h*-polynomial h*(P;u,v), whose top-degree part is the local h*-polynomial ℓ*(P;u,v). The proof relies on a closed formula (Prop. 2.7) expressing the Hodge-Deligne polynomial of a generic complete intersection in the algebraic torus as a sum over subsets I of the polytopes, involving h*(P_I;u,v) and powers of (uv-1). The Cayley polytope P1*...*Pk embeds the complete intersection as a hypersurface, and the formula lets the authors isolate the top-degree term. The mixed volume MV(P1,...,Pk) appears as the number of solutions to the first k equations (BKK theorem), counted fiberwise over the projection along U.

What would settle it

Take k=1, P1 the unit interval along the x-axis in R^2, and Q a triangle with vertices (0,0),(2,0),(0,1), so U is the x-axis and proj_U(Q) has normalized length 2. Compute both sides of ℓ*(P1*Q;u,v) = 1·uv·ℓ*(proj_U(Q);u,v) using the explicit combinatorial formulas for 3- and 2-dimensional polytopes; any discrepancy would disprove Theorem 1.2. More directly, search for a coefficient vector where a specialization g_{x'}(z) is not Newton-regular; if for such a choice the Hodge-Deligne polynomial of the fiber differs from the uniform expression (2.4), the fiberwise step in the proof fails, though

Watch

Extended reading notes

Core claim

Let P1,...,Pk,P_{k+1} be lattice polytopes in R^d with P1,...,Pk contained in a k-dimensional rational subspace U and dim P_{k+1}=d. The paper proves that the bivariate local h*-polynomial of the Cayley polytope P1*...*P_{k+1} satisfies ℓ*(P_{[k+1]};u,v) = MV(P1,...,Pk) (uv)^k ℓ*(proj_U(P_{k+1});u,v), so their Hodge vectors agree up to the scalar MV(P1,...,Pk). The argument equates two expressions for the Hodge-Deligne polynomial of the complete intersection defined by the associated Laurent polynomials: one from a closed formula for complete intersections, the other by stratifying the solution set into fibers over the finite zero-dimensional set of the first k equations, whose size is the m

Load-bearing premise

The argument assumes that the generic Laurent polynomial defining the last polytope and all of its specializations to the finite zero set of the first k polynomials are simultaneously regular with respect to their Newton polytopes; if this simultaneous regularity fails on a positive-dimensional subset of the coefficients, the fiberwise Hodge-Deligne computation would no longer be uniform.

Editorial extensions

If this is right

For any lattice polytope P, one can build infinitely many non-isomorphic polytopes in dimensions dim P + 2k (k≥1) with the same Hodge vector, none of which are free joins.
The question of whether every thin spanning polytope that is not a free join must be trivially thin is answered negatively in every dimension ≥5: Lawrence twists produce infinitely many counterexamples.
B_k-polytopes are thin for a simple reason: they are generalized Lawrence twists with mixed volume zero.
Free joins with a lattice interval of length 1 satisfy an explicit h*-polynomial relation h*(P*I;t) = t h*(proj_I P,t) + h*(P,t), giving degree growth under projection.
Every odd dimension ≥3 contains infinitely many non-isomorphic nearly thin polytopes (Hodge vector (1)) that are not free joins.

Reading between the lines

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

If the formula holds, the Hodge vector of the Cayley polytope depends on the first k factors only through their mixed volume, not their individual shapes — a kind of universality that may reflect an underlying toric fibration whose monodromy is governed by the mixed volume.
The construction suggests a practical route to test dimensional reduction of hypergeometric motives: any motive realized by a hypersurface with Hodge vector of length ℓ might also be realized by a lower-dimensional hypersurface obtained by projection, at least within the toric family.
The h*-polynomial relation for P*I could lead to new bounds on the degree of projections of lattice polytopes, and possibly to an algorithm for computing h* of a projection recursively.
The quotient structure suggests a congruence relation on polytopes generated by Cayley products with mixed volume 1; one could try to classify such equivalence classes.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Desk Editor's Note

read the letter

Main takeaway: the paper's central formula (Theorem 3.1) is new and I believe correct. It computes the bivariate h*-polynomial of a Cayley polytope P1*...*Pk+1 as mixed volume times projection plus a correction sum, and under the stated conditions the top-degree part collapses to V(uv)^k times the bivariate local h*-polynomial of the projection. That is a genuine generalization of the free-join/interval case and gives a clean way to build infinite families with the same Hodge vector that are not free joins. The proof via Proposition 2.7 is coherent: the Danilov–Khovanskii stratification and (2.4) do lead to (3.1) after cancellation; I did not find a fatal gap.

The weakest part is not the algebra but the regularity footnote. The fiberwise argument needs the specialized polynomials g_{x'}(z) to be simultaneously Newton-regular. The footnote's justification is too quick. That said, I think the concern is repairable: for each fixed choice of f1..fk, the bad c form a finite union of hyperplanes, and a dimension count over the product coefficient space gives a Zariski open set of good pairs. The paper should spell this out, because as written it is the least secure step.

More significant for the applications: Lemma 5.1, which supplies infinitely many non-isomorphic non-free-join Lawrence twists, is stated with the proof deferred to the first author's thesis. That is load-bearing for the advertised infinitude. It should be proven in the paper or the thesis argument included. Corollary 4.4 is also credited to the thesis, but that one is less central. Proposition 2.6 and Remark 3.3 being left as exercises is a minor style issue, not a problem.

The citation pattern is fine. The self-citations are to published/thesis results, and the reliance on [DRHN19] for the stratification is appropriate. No sign of circularity.

Who this is for: people working in local Ehrhart theory, thin polytopes, and hypergeometric motives. It answers questions from BKN23 and Selyanin and offers tools that should be used.

Recommendation: send to peer review. A referee should push for a full proof of Lemma 5.1 and a rigorous regularity argument in Theorem 3.1, but the central result deserves to be in the literature.

Referee Report

3 major / 5 minor

Summary. The paper proves a formula for the bivariate h*-polynomial of a Cayley polytope P_1 * ... * P_{k+1} when the first k polytopes lie in a k-dimensional subspace U; the leading term is the mixed volume V = MV(P_1,...,P_k) times the bivariate h*-polynomial of the projection of P_{k+1} along U. From this it derives equality of Hodge vectors (local h*-vectors with zeroes removed) and uses it to construct infinitely many non-isomorphic polytopes with the same Hodge vector that are not free joins, as well as new thin and nearly thin polytopes. The main tool is a closed formula for the Hodge-Deligne polynomial of a generic complete intersection in the torus in terms of bivariate h*-polynomials of Cayley polytopes. The paper also introduces Lawrence twists and relates them to Gale duality.

Significance. If the proof is fully rigorous, the result is significant: it gives the first general mechanism producing infinitely many high-dimensional lattice polytopes with the same Hodge vector outside the free-join construction, and it answers a question on thin polytopes. The paper is written in a transparent, example-rich style and builds on standard external results (Danilov--Khovanskii, BKK, bivariate Ehrhart theory); there is no circularity in the central derivation. The constructions are explicit and testable, and the applications to thin polytopes and B_k-polytopes are concrete.

major comments (3)

[§3, Theorem 3.1 proof, footnote 2] The claim that all restricted polynomials g_{x'}(z) = f_{k+1}(x',z) are simultaneously regular is not rigorously justified. The finite set bY depends on f_1,...,f_k, so the bad sets for x' cannot simply be treated as independent hypersurfaces in the coefficient space of f_{k+1}; x' moves with the coefficients of f_1,...,f_k. A repair is to choose f_1,...,f_k generically first, then choose the coefficients of f_{k+1} outside the finite union of hypersurfaces determined by the resulting finite set bY; one must also ensure that this choice is compatible with the Zariski-open condition required by Proposition 2.7. Please provide this argument or a direct elimination-based proof of simultaneous regularity.
[§3, Theorem 3.1 proof, case k>d] The sentence 'For the case when k > d, the equations (3.3) and (3.4) that one derives from Proposition 2.7 still hold with V=0' is false as stated. Equation (3.3) was derived under the assumption k≤d, where f_1,...,f_k are restricted to (C*)^k. For k>d that restriction is impossible, and the literal identity (3.3) with V=0 is not a consequence of Proposition 2.7. For example, for d=1, k=2, P_1=P_2=[0,1], the right-hand side of (3.3) with V=0 equals (u^2v^2 - uv - 2)/(uv)^2, not 0. The formula (3.1) in the case k>d should be derived directly from Proposition 2.7 applied to the overdetermined system, using that the sum over I⊆[k] vanishes. Please rewrite this part of the proof.
[§5.1, Lemma 5.1] Lemma 5.1 is a load-bearing statement for the applications (Corollaries 5.4 and 5.7), but its proof is left to the reader and deferred to the first author's thesis [Kur24a]. For a journal article, this is not sufficient: either include a complete proof in the paper or give a precise reference to a published/freely accessible proof that the reader can verify. The remark following the lemma indicates that the proof is subtle at the level of point configurations, so deferring it is particularly problematic.

minor comments (5)

[§2.3, Proposition 2.6] Typo: 'Is is straightforward' should be 'It is straightforward'. Also, the proof is left to the reader; a citation to [NS13, Remark 4.6(5)] would be helpful.
[§4.2, Definition 4.2] The notation 'ĕA ⊂ Z^{d+2k}' is slightly imprecise: ĕA is a point configuration, not a subset of the lattice in the usual sense. Please phrase as 'an integer point configuration in Z^{d+2k}'.
[§5.6] The statement 'The reader is invited to check that in dimension 2 every thin polytope is a generalized Lawrence twist except for 2Δ_2' is an unproved classification claim. If it is known, provide a reference; otherwise phrase it as a conjecture or give the short proof.
[Abstract and Theorem 1.2] The abstract uses 'P' for the last polytope while Theorem 1.2 uses P_{k+1}; unify the notation.
[References] The spelling 'Kouchnirenko' is inconsistent with the Cyrillic transliteration used elsewhere; please standardize.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: the Hodge-vector formula is derived from independent external results, not from a fitted input or a self-referential definition.

full rationale

The derivation of Theorem 3.1 / Theorem 1.2 is not circular. Proposition 2.7 is proved in the paper from the Danilov–Khovanskii formula (2.7), the stratification from [DRHN19], and the bivariate Ehrhart identity (2.4); none of these inputs contains the target equality between the Cayley polytope's ℓ*-polynomial and the projection's ℓ*-polynomial. Equation (3.3) is an application of Proposition 2.7 and the BKK theorem, identifying the mixed volume V with the size of the zero-dimensional complete intersection; V is not a fitted value. Equation (3.5) is a separate fiberwise computation over the finite set bY using regularity of each specialized Laurent polynomial. Footnote 2's simultaneous-regularity assertion is a standard Zariski-open avoidance argument over finitely many hypersurface conditions and does not presuppose the Hodge-vector equality. Equating (3.4) and (3.5) and taking top-degree parts yields (3.1)/(3.2) by algebra. The main self-citations — [DRHN19] for a stratification, [Kur24a] for Lemma 5.1 and Lawrence-twist construction — are independent prior results or are accompanied by proofs/constructions in this paper; they are not fitted parameters and do not force the central claim by definition. The only point worth noting is the terseness of the simultaneous-regularity genericity claim, which is a technical-rigor concern rather than a circularity.

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

The paper defines new combinatorial constructions ('Lawrence twist', 'generalized Lawrence twist') but these are definitions, not postulated entities with independent empirical handles; no new particles, forces, or dimensions are introduced. No free parameters were fitted: mixed volumes and projections are computed, not tuned.

assumptions (6)

standard math Danilov-Khovanskii formula (2.7) for the E-polynomial of the universal hypersurface, and Proposition 2.1 on primitive cohomology.
Used in the proof of Prop 2.7 and throughout Section 2; taken from [DK87].
standard math Bernstein-Khovanskii-Kushnirenko theorem: number of isolated solutions of a generic sparse system equals mixed volume of Newton polytopes.
Used in Section 3 to identify |Y_hat| with MV(P1,...,Pk) and to justify V=0 in the overdetermined case.
standard math Katz-Stapledon bivariate h*-polynomial formalism and its relation to Hodge-Deligne polynomials (Eqs. 2.3, 2.4).
Defines h*(P;u,v), ell*(P;u,v) and the Hodge vector; assumed as background from [KS16, BM03].
domain assumption Generic regularity: there is a Zariski-open set of coefficient vectors for which f_{k+1} and each specialized fiber g_{x'}(z) are simultaneously regular with respect to their Newton polytopes.
Stated in footnote 2; if simultaneous regularity failed, the fiberwise E-polynomial computation (3.5) would not hold.
standard math Any rational k-dimensional subspace U can be mapped by a unimodular transformation to the coordinate subspace R^k x {0}.
Used at the start of the proof of Theorem 3.1 to reduce to f1,...,fk depending only on x1,...,xk.
domain assumption Gale duality correspondence between spanning point configurations and vector configurations, and the claim that Lawrence twists are non-isomorphic and not free joins.
Underpins Lemma 5.1 and the infinite-family claims; proof of Lemma 5.1 is deferred to [Kur24a].

how reviews work

0 comments

Cite this review

Pith. "Pith review of Preserving Hodge Vectors of Lattice Polytopes." pith.science (2026). https://pith.science/paper/DO7KQPAS

@misc{pith2026260220765,
  author       = {Pith},
  title        = {Pith review of: Preserving Hodge Vectors of Lattice Polytopes},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DO7KQPAS}},
  note         = {Machine review of arXiv:2602.20765}
}

abstract

Given lattice polytopes $P_1, \ldots, P_k$ contained in a $k$-dimensional subspace $U \subseteq \mathbb{R}^d$ and a $d$-dimensional lattice polytope $Q \subset \mathbb{R}^d$, we compute the Hodge vector of the Cayley polytope $P_1 * \cdots * P_k * Q$, and show that it equals the mixed volume of $P_1, \ldots, P_k$ times the Hodge vector of the projection of $Q$ along $U$. Here, the Hodge vector of a lattice polytope is its local $h^*$-vector with leading and trailing zeroes removed. This result allows finding infinitely many high-dimensional lattice polytopes with the same Hodge vector that are not free joins. The proof relies on a closed formula for the Hodge-Deligne polynomial of generic complete intersections in the torus in terms of the bivariate/mixed $h^*$-polynomial. A special case of our construction is what we call Lawrence twists: extending the Gale transform by centrally-symmetric pairs of vectors. As applications, we can produce many new thin polytopes answering a question by Borger, Kretschmer and the second author, and we provide an alternative explanation of the thinness of $B_k$-polytopes answering a question of Selyanin.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

4 extracted references · 1 linked inside Pith

[4]

Newton numbers, vanishing polytopes and algebraic degrees

[Sel25] Fedor Selyanin. Newton numbers, vanishing polytopes and algebraic degrees. Preprint, arXiv:2507.03661,

arXiv
[1993]

[Bat06] Victor V

Publisher: Duke University Press. [Bat06] Victor V. Batyrev. Lattice polytopes with a givenh ∗-polynomial. InAlgebraic and geometric com- binatorics. Euroconference in mathematics, algebraic and geometric combinatorics, Anogia, Crete, Greece, August 20–26, 2005, pages 1–10. Providence, RI: American Mathematical Society (AMS),

2005
[1994]

Hypergeometric local systems overQwith Hodge vector (1,1,1,1)

[GV24] Giulia Gugiatti and Fernando Rodriguez Villegas. Hypergeometric local systems overQwith Hodge vector (1,1,1,1). Preprint, arXiv:2401.13529,

arXiv
[2024]

Thin Simplices via Modular Arithmetic

[Kur24b] Vadym Kurylenko. Thin Simplices via Modular Arithmetic. Preprint, arXiv:2404.03975,

arXiv

Pith tools

Reviewed August 2, 2026 · model on record in the stance chip above.

[1] [4]

Newton numbers, vanishing polytopes and algebraic degrees

[Sel25] Fedor Selyanin. Newton numbers, vanishing polytopes and algebraic degrees. Preprint, arXiv:2507.03661,

arXiv

[2] [1993]

[Bat06] Victor V

Publisher: Duke University Press. [Bat06] Victor V. Batyrev. Lattice polytopes with a givenh ∗-polynomial. InAlgebraic and geometric com- binatorics. Euroconference in mathematics, algebraic and geometric combinatorics, Anogia, Crete, Greece, August 20–26, 2005, pages 1–10. Providence, RI: American Mathematical Society (AMS),

2005

[3] [1994]

Hypergeometric local systems overQwith Hodge vector (1,1,1,1)

[GV24] Giulia Gugiatti and Fernando Rodriguez Villegas. Hypergeometric local systems overQwith Hodge vector (1,1,1,1). Preprint, arXiv:2401.13529,

arXiv

[4] [2024]

Thin Simplices via Modular Arithmetic

[Kur24b] Vadym Kurylenko. Thin Simplices via Modular Arithmetic. Preprint, arXiv:2404.03975,

arXiv