Makhlin polytopes are Demazure string polytopes

Ghislain Fourier

Makhlin polytopes are Demazure string polytopes

T0 review · 0 major / 3 minor · reviewed 2026-08-28 · deepseek-v4-flash

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

Pith's one-line read Makhlin's type-$B$ polytopes are affinely unimodularly equivalent to Demazure string polytopes in type $B_{2n-1}$, with an equivalence compatible with addition of weights.

desk verdict A clean composition proof that Makhlin's type B polytopes are unimodularly equivalent to Demazure string polytopes in type B_{2n-1}; the new cancellation of diagonal factors is genuine, though the proof leans on an imported coordinate identification from [CFL24]. read the letter →

arxiv 2608.23431 v1 pith:5L2TLOGS submitted 2026-08-24 math.CO math.RT

Ghislain Fourier This is my paper · ORCID

classification math.COmath.RT MSC 17B1017B3752B2014M25

keywords MakhlinpolytopeDemazurestringFFLVtypeBconePBWfiltrationunimodularequivalencelattice-pointsemigroup

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 that Makhlin's polytopes, the convex models that parametrize monomial bases for the PBW filtration in type $B_n$, are not new polyhedral objects: each one is affinely unimodularly equivalent to a Demazure string polytope for a higher-rank root system of the same type. Concretely, for every dominant integral weight $\lambda = \sum a_i \omega_i$ of type $B_n$, the Makhlin polytope $P^{B_n}_{\mathrm{Mak}}(\lambda)$ is mapped bijectively onto the string polytope $Q^{B_{2n-1}}_{w_n}(\tilde\lambda)$, where $\tilde\lambda = \sum a_i \tilde\omega_{2i-1}$ and $w_n$ is a fixed reduced word. The equivalence can be chosen compatibly with addition of dominant weights, so it upgrades to an isomorphism of the multigraded lattice-point semigroups. This matters because types $A$ and $C$ already had such a PBW-to-string shadow, and the present argument extends it to type $B$ by composing three existing maps rather than repeating the full analysis. The takeaway is that the lattice points counting these bases are the same combinatorial objects as Demazure string bases in a higher rank.

What carries the argument

The mechanism is the identity $\Gamma_n = 2D_n^{-1}$ (Lemma 3.3) in the common coordinate system of [CFL24, Notation 4.10], where the same root labels index both the symplectic FFLV coordinates and the string coordinates of the word $w_n$. Makhlin's comparison map halves exactly those long-root boundary coordinates that the $B/C$ string-cone similarity doubles, so the two non-unimodular maps cancel when composed in opposite directions. What remains is the linear part $U_n = D_n X_n D_n^{-1}$, which is integral and unimodular because it is a conjugate of the integral unimodular map $X_n$, and the translation part $c_\lambda = \frac12 D_n t_{n,\mu}$, which is integral by Lemma 2.2 because the coefficients on the long-root boundary coordinates of $t_{n,\omega_n}$ are all $1$. This cancellation is the whole point of passing through type $C$.

What would settle it

Compute the lattice-point counts (or Ehrhart series) of $P^{B_n}_{\mathrm{Mak}}(\lambda)$ and $Q^{B_{2n-1}}_{w_n}(\tilde\lambda)$ for a small case such as $n=3$ with $\lambda = \omega_1 + \omega_2 + \omega_3$; any mismatch in the number of integral points, or any non-integral value of $\Phi_{n,\lambda}$ on a vertex, would falsify the theorem. A more targeted check is to verify the inequality list of Example 3.2: if the image of the type-$B$ polytope under the composed map is not cut out by exactly the displayed type-$C$ inequalities, the $B/C$ similarity step for this word fails.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central claim is Theorem 1.1. Let $\lambda = \sum_{i=1}^n a_i \omega_i^{B_n}$ be a dominant integral weight of type $B_n$, and set $\tilde\lambda = \sum_{i=1}^n a_i \tilde\omega_{2i-1}$ in type $B_{2n-1}$. Then Makhlin's polytope $P^{B_n}_{\mathrm{Mak}}(\lambda)$ is affinely unimodularly equivalent to the Demazure string polytope $Q^{B_{2n-1}}_{w_n}(\tilde\lambda)$, where $w_n = \tau_m \tau_{m-1} \cdots \tau_{n+1} \sigma_n \cdots \sigma_1$ is the reduced word used in the type-$C$ construction. The equivalence is compatible with addition of dominant weights and, at the level of integral points, sends Makhlin's monomial basis to the Demazure string basis. The proof realizes the equivalence as the composition $\Phi_{n,\lambda} = \Gamma_n^{-1} \circ T_{n,\mu} \circ 2D_n^{-1}$, in which the two non-unimodular diagonal rescalings cancel.

Load-bearing premise

The load-bearing premise is that the same root labels can serve simultaneously as coordinates for the symplectic FFLV polytope and for the string coordinates of the fixed word $w_n$; if that shared labelling is not valid for this word, the identity $\Gamma_n = 2D_n^{-1}$ and the integrality of the composed map collapse.

Editorial extensions

If this is right

For every dominant integral weight $\lambda$ of type $B_n$, the lattice points of the Makhlin polytope are in bijection with those of the higher-rank Demazure string polytope, so Makhlin's monomial basis is a Demazure string basis in type $B_{2n-1}$.
The compatibility with addition implies an isomorphism of multigraded lattice-point semigroups $\Sigma_{\mathrm{Mak}} \cong \Sigma_{\mathrm{Str}}$ (Corollary 5.1).
The string-side semigroup is generated in degrees $\omega_1, \ldots, \omega_n, 2\omega_n$; equivalently by the lattice-point sets $S_{\mathrm{Str}}(\tilde\omega_1), \ldots, S_{\mathrm{Str}}(\tilde\omega_{2n-1}), S_{\mathrm{Str}}(2\tilde\omega_{2n-1})$ (Corollary 5.2).
The type-$B$ equivalence factors through the symplectic FFLV polytopes and the string-cone similarity, so type-$B$ PBW polytopes inherit the toric-degeneration structure of the higher-rank Demazure polytopes without repeating the type-$A$/$C$ analysis.

Reading between the lines

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

Because the proof only uses the three maps and the cancellation of two diagonal rescalings, the same mechanism may apply to other PBW-type polytopes, such as the $G_2$ model mentioned in the paper, whenever a diagonal comparison map and a string-cone similarity are both available.
The explicit form $\Phi_{n,\lambda}(q) = D_n X_n D_n^{-1} q + \tfrac12 D_n t_{n,\mu}$ gives an algorithm to convert Makhlin monomials into Demazure string basis elements; one could implement it to compute transition matrices or compare with other monomial bases for small $n$.
The appearance of the degree $2\omega_n$ generator on the string side suggests that Makhlin's splitting off of $2\omega_n$ is not a normalization artifact but encodes how the half-root weight $\omega_n^{B_n}$ is embedded into the higher-rank even root lattice, possibly with a representation-theoretic reading in the Demazure module.
The proof depends on the fixed word $w_n$; a testable extension is whether the equivalence remains unimodular for other reduced words related by commutation or braid moves, or whether $w_n$ is essentially forced by the diagonal comparison.

Share X Bluesky LinkedIn Reddit HN

Formalized claims in Lean

Claim #1: On the paper's own terms, the central claim is Theorem 1.1. Let $\lambda = \sum_{i=1}^n a_i \omega_i^{B_n}$ be a dominant integral weight of type $B_n$, and set $\tilde\lambda = \sum_{i=1}^n a_i \tilde\omega_{2i-1}$ in type $B_{2n-1}$. Then Makhlin's polytope $P^{B_n}_{\mathrm{Mak}}(\lambda)$ is affinely unimodularly equivalent to the Demazure string polytope $Q^{B_{2n-1}}_{w_n}(\tilde\lambda)$, w

/-- @claim 1 On the paper's own terms, the central claim is Theorem 1.1. Let $\lambda = \sum_{i=1}^n a_i \omega_i^{B_n}$ be a dominant integral weight of type $B_n$, and set $\tilde\lambda = \sum_{i=1}^n a_i \tilde\omega_{2i-1}$ in type $B_{2n-1}$. Then Makhlin's polytope $P^{B_n}_{\mathrm{Mak}}(\lambda)$ is affinely unimodularly equivalent to the Demazure string polytope $Q^{B_{2n-1}}_{w_n}(\tilde\lambda)$, w -/
def central_claim : Prop :=

Signed reviews

No signed human review yet.

Request a human review

A listed scientist reviews the paper for a fee and the review publishes here regardless of verdict. See the reviewers or get listed.

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

First thing to know: the paper proves that Makhlin's type B polytopes are unimodularly equivalent to Demazure string polytopes in type B_{2n-1}, by composing three known maps. The genuinely new step is a cancellation: the non-unimodular diagonal rescaling in Makhlin's comparison with symplectic FFLV polytopes is exactly the inverse of the B/C similarity on the string side, so in the composition they cancel and leave an integral unimodular map.

The paper does that cleanly. Lemma 3.3 (Γ_n = 2D_n^{-1}) is the heart, and it's a real observation, not a routine check. Proposition 3.1 extends the B/C similarity from string cones to highest-weight cuts; the proof is direct. Propositions 4.1 and 4.2 verify integrality of the linear and translation parts explicitly. The example n=3 is helpful, and the semigroup corollaries are a nice consequence. The writing is honest about what is imported and what is new.

The soft spots, in proportion: the proof leans on the coordinate identification from [CFL24, Notation 4.10], namely that the same root labels index the FFLV coordinates and the string coordinates of the word w_n. That identification is load-bearing for Lemma 3.3 and for the conjugation U_n = D_n X_n D_n^{-1}. It is quoted, not proved, in this paper. If that convention were wrong for w_n, the cancellation and integrality would collapse. I see no reason to think it fails, but a referee should check that citation page carefully. Lemma 2.2 is also terse: one sentence about coefficients of t_{n,ω_n}. The claim is concrete and checkable, so I don't consider it a serious gap. The heavier dependence on [CFL24] and [CFL25] as black boxes is standard in this area; one of those is the author's own group, but the citation is appropriate and not self-promoting.

My verdict: the central argument holds up if you accept the black boxes. The new cancellation is real and the paper fills a genuine gap in type B. This deserves a serious referee, not a desk reject. I'd send it to someone who knows string polytopes and PBW monomial bases, with a specific instruction to verify the coordinate identification from [CFL24] and to redo Lemma 2.2. For readers working on toric degenerations or PBW filtration, this is a useful and citable result.

Referee Report

0 major / 3 minor

Summary. This paper proves that Makhlin's polytopes for type B_n are affinely unimodularly equivalent to Demazure string polytopes for a specific reduced word in type B_{2n-1} (Theorem 1.1), with the equivalence compatible with addition of weights. The proof composes three existing maps: Makhlin's diagonal comparison to symplectic FFLV polytopes (Proposition 2.3), the affine unimodular equivalence of [CFL24] (Theorem 2.1), and the inverse of the B/C similarity of string cones from [CFL25] (Proposition 3.1). The new contributions are the extension of the B/C similarity to highest-weight cuts, the identification Gamma_n = 2D_n^{-1} (Lemma 3.3), and the integrality of the resulting linear part and translation (Propositions 4.1 and 4.2). Section 5 derives consequences for lattice-point semigroups and their generation.

Significance. If correct, Theorem 1.1 provides the combinatorial shadow of the PBW filtration for type B, a case where no ordinary FFLV polytope exists, and connects Makhlin's polytopes to the well-studied theory of Demazure string polytopes. The proof is elegant and mostly composed of previously established results; the new steps are elementary but necessary. The paper is transparent about its reliance on [CFL24], including the coordinate identification used throughout. The explicit formulas for U_n and c_lambda make the construction checkable, and the additivity result yields a clean semigroup isomorphism.

minor comments (3)

[Section 2.2, Lemma 3.3] The proof of Lemma 3.3 depends on the imported property iota(a,b)=m iff b=bar{a} from [CFL24, Notation 4.10] and on the identification of the FFLV and string coordinates of w_n with one common R^N; since this identification is load-bearing for the conjugation U_n = D_n X_n D_n^{-1} and for the cancellation of the diagonal rescalings, I recommend that the authors restate the identification explicitly (for example, by writing out the bijection between the coordinates e_{a,b} and the positions of the simple reflections in w_n) to make the paper more self-contained.
[Lemma 2.2] The proof of Lemma 2.2 is very terse ('the first case is empty'); adding a few words on how [CFL24, Definition 4.5] yields the stated coefficients would improve readability.
[Section 2.1] The sentence 'We use the standard lattice in string coordinates' would benefit from a brief explanation that the lattice is Z^L in the given basis, so that the unimodularity statements are unambiguous.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: Theorem 1.1 composes three published black-box equivalences, with the new work confined to integrality and highest-weight checks.

full rationale

Theorem 1.1 is not derived from itself. The proof explicitly composes three previously established maps: Makhlin's diagonal map 2D_n^{-1} (Proposition 2.3, from [Mak19, Remark 2.3]), the type C FFLV-to-string equivalence T_{n,mu} (Theorem 2.1, from [CFL24]), and the inverse B/C similarity Gamma_n^{-1} (Proposition 3.1, from [CFL25]). None of these inputs contains the target statement that Makhlin's type B_n polytope is unimodularly equivalent to the Demazure string polytope in type B_{2n-1}; the target result is assembled, not assumed. The only genuinely new steps are verifying that Gamma_n^{-1} preserves the highest-weight inequalities, establishing the cancellation identity Gamma_n = 2D_n^{-1} (Lemma 3.3), and proving integrality of U_n and c_lambda (Propositions 4.1 and 4.2). These checks are performed directly on displayed formulas and do not fit any parameter to the claimed output. The reliance on [CFL24, Notation 4.10] to identify FFLV and string coordinates in one common R^N is an imported convention rather than a redefinition of the theorem; even if that convention failed, the issue would be a correctness gap, not circularity. The self-citation [CFL24] is a published theorem with stated assumptions that do not include the target result, so under the given rules it counts as independent evidence and does not raise the circularity score. No equation in the paper reduces to another by construction, and no fitted quantity is renamed as a prediction.

Assumptions & free parameters 0 free parameters · 7 assumptions · 0 invented entities

No fitted constants or empirical data appear. The paper is a composition of previously established theorems: Makhlin's diagonal comparison, the type C FFLV-to-string equivalence, and the B/C string-cone similarity, plus a common coordinate identification imported from [CFL24]. These are recorded as domain assumptions. No new particles, forces, or other entities are introduced.

assumptions (7)

domain assumption Theorem 2.1 from [CFL24]: affine map T_{n,mu}(x)=X_n x + t_{n,mu} with X_n in GL_N(Z) maps the type C_n FFLV polytope to the string polytope Q^C_m_wn(tilde_mu).
Imported black box; provides the unimodular middle step of the composition.
domain assumption Theorem 3.2 from [CFL25]: the diagonal map Gamma^{B,C} identifies type B_m and C_m string cones for every reduced word of the longest element.
Imported black box; Proposition 3.1 extends it to highest-weight cuts.
domain assumption Remark 2.3 from [Mak19]: Makhlin's type B polytope is diagonally related to the type C FFLV polytope; restated as PBn_Mak(lambda)=1/2 D_n PCn_FFLV(mu) in Proposition 2.3.
Imported comparison; supplies the first map in the diagram.
domain assumption Notation 4.10 from [CFL24]: the FFLV coordinates and the string coordinates of w_n are indexed by the same root labels.
Load-bearing for the conjugation U_n = D_n X_n D_n^{-1}; without it the cancellation is not meaningful.
standard math The word w_n = tau_m ... tau_{n+1} sigma_n ... sigma_1 is reduced of length N=n^2 in the common Weyl group W(B_m)=W(C_m).
Verified by explicit length count in Section 2.1; needed for string polytopes to be well-defined.
standard math Highest-weight inequalities for string polytopes have the form (2) with the Cartan matrix A=(a_{ij}).
Quoted from Littelmann [Lit98] and Berenstein-Zelevinsky [BZ01]; used in Proposition 3.1.
domain assumption Positive roots of B_n and C_n are identified by epsilon_i-epsilon_j maps to epsilon_i-epsilon_j, epsilon_i+epsilon_j maps to epsilon_i+epsilon_j, and epsilon_i maps to 2epsilon_i.
Makes the Dyck path combinatorics of Makhlin and FFLV polytopes correspond; used in Proposition 2.3.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Makhlin polytopes are Demazure string polytopes." pith.science (2026). https://pith.science/paper/5L2TLOGS

@misc{pith2026260823431,
  author       = {Pith},
  title        = {Pith review of: Makhlin polytopes are Demazure string polytopes},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/5L2TLOGS}},
  note         = {Machine review of arXiv:2608.23431}
}

abstract

We show that Makhlin's polytopes in type $B_n$ are unimodularly equivalent to string polytopes for Demazure modules in type $B_{2n-1}$. The proof passes through type $C$, combining Makhlin's diagonal comparison with symplectic FFLV polytopes and the $B/C$ similarity for string cones.

Figures

Figures reproduced from arXiv: 2608.23431 by the authors.

**Figure 1.** The three maps used in the proof. Let wn be the reduced word used in the type C construction of [CFL24]; it is recalled explicitly in Section 2. The main result is the following. Theorem 1.1. Let λ = Pn i=1 aiωi be a dominant integral weight of type Bn and set λe = Xn i=1 aiωe2i−1 in type B2n−1. Then Makhlin’s polytope P Bn Mak(λ) is affinely unimodularly equivalent to the Demazure string polytope Q B2n−1 wn (λe). T… view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

22 extracted references · 22 canonical work pages

[1]

Inventiones Mathematicae , volume =

Berenstein, Arkady and Zelevinsky, Andrei , title =. Inventiones Mathematicae , volume =. 2001 , pages =

work page 2001
[2]

Transformation Groups , volume =

Caldero, Philippe , title =. Transformation Groups , volume =. 2002 , pages =

work page 2002
[3]

Mathematische Zeitschrift , volume =

Cerulli Irelli, Giovanni and Fang, Xin and Feigin, Evgeny and Fourier, Ghislain and Reineke, Markus , title =. Mathematische Zeitschrift , volume =. 2017 , pages =

work page 2017
[4]

2024 , eprint =

Cleusters, Ester and Fourier, Ghislain and Lerner, Felix , title =. 2024 , eprint =

work page 2024
[5]

Dvořáková, K

Cho, Yunhyung and Fujita, Naoki and Lee, Eunjeong , title =. European Journal of Combinatorics , volume =. 2025 , pages =. doi:10.1016/j.ejc.2025.104126 , eprint =

work page arXiv 2025
[6]

International Mathematics Research Notices , number =

Cerulli Irelli, Giovanni and Lanini, Martina , title =. International Mathematics Research Notices , number =. 2015 , pages =

work page 2015
[7]

Pacific Journal of Mathematics , volume =

Cerulli Irelli, Giovanni and Lanini, Martina and Littelmann, Peter , title =. Pacific Journal of Mathematics , volume =. 2016 , pages =

work page 2016
[8]

2024 , eprint =

Enugandla, Shreepranav Varma and Fang, Xin and Fourier, Ghislain and Steinert, Christian , title =. 2024 , eprint =

work page 2024

Show all 22 references

[9]

Mathematical Research Letters , volume =

Feigin, Evgeny , title =. Mathematical Research Letters , volume =. 2011 , pages =

2011
[10]

Selecta Mathematica (N.S.) , volume =

Feigin, Evgeny , title =. Selecta Mathematica (N.S.) , volume =. 2012 , pages =

2012
[11]

Transformation Groups , volume =

Feigin, Evgeny and Fourier, Ghislain and Littelmann, Peter , title =. Transformation Groups , volume =. 2011 , pages =

2011
[12]

International Mathematics Research Notices , number =

Feigin, Evgeny and Fourier, Ghislain and Littelmann, Peter , title =. International Mathematics Research Notices , number =. 2011 , pages =

2011
[13]

Canadian Journal of Mathematics , volume =

Feigin, Evgeny and Finkelberg, Michael and Littelmann, Peter , title =. Canadian Journal of Mathematics , volume =. 2014 , pages =

2014
[14]

Lie Algebras and Their Representations , series =

Kashiwara, Masaki , title =. Lie Algebras and Their Representations , series =. 1996 , pages =

1996
[15]

Transformation Groups , volume =

Littelmann, Peter , title =. Transformation Groups , volume =. 1998 , pages =

1998
[16]

Algebraic Combinatorics , volume =

Makhlin, Igor , title =. Algebraic Combinatorics , volume =. 2019 , pages =

2019
[17]

Gornitskii, A. A. , title =. Mathematical Notes , volume =. 2015 , pages =

2015
[18]

Backhaus, Teodor and Ku. The. Journal of Pure and Applied Algebra , volume =. 2019 , pages =

2019
[19]

2022 , eprint =

Enugandla, Shreepranav Varma , title =. 2022 , eprint =

2022
[20]

Combinatorial Theory , volume =

Makedonskyi, Ievgen and Makhlin, Igor , title =. Combinatorial Theory , volume =. 2025 , eid =

2025
[21]

Journal of Combinatorial Theory, Series A , volume =

Fourier, Ghislain , title =. Journal of Combinatorial Theory, Series A , volume =. 2016 , pages =

2016
[22]

Beitr\"age zur Algebra und Geometrie / Contributions to Algebra and Geometry , volume =

Balla, George and Fourier, Ghislain and Kambaso, Kunda , title =. Beitr\"age zur Algebra und Geometrie / Contributions to Algebra and Geometry , volume =. 2023 , pages =

2023

Pith tools

Reviewed August 28, 2026 · model on record in the stance chip above.

[1] [1]

Inventiones Mathematicae , volume =

Berenstein, Arkady and Zelevinsky, Andrei , title =. Inventiones Mathematicae , volume =. 2001 , pages =

work page 2001

[2] [2]

Transformation Groups , volume =

Caldero, Philippe , title =. Transformation Groups , volume =. 2002 , pages =

work page 2002

[3] [3]

Mathematische Zeitschrift , volume =

Cerulli Irelli, Giovanni and Fang, Xin and Feigin, Evgeny and Fourier, Ghislain and Reineke, Markus , title =. Mathematische Zeitschrift , volume =. 2017 , pages =

work page 2017

[4] [4]

2024 , eprint =

Cleusters, Ester and Fourier, Ghislain and Lerner, Felix , title =. 2024 , eprint =

work page 2024

[5] [5]

Dvořáková, K

Cho, Yunhyung and Fujita, Naoki and Lee, Eunjeong , title =. European Journal of Combinatorics , volume =. 2025 , pages =. doi:10.1016/j.ejc.2025.104126 , eprint =

work page arXiv 2025

[6] [6]

International Mathematics Research Notices , number =

Cerulli Irelli, Giovanni and Lanini, Martina , title =. International Mathematics Research Notices , number =. 2015 , pages =

work page 2015

[7] [7]

Pacific Journal of Mathematics , volume =

Cerulli Irelli, Giovanni and Lanini, Martina and Littelmann, Peter , title =. Pacific Journal of Mathematics , volume =. 2016 , pages =

work page 2016

[8] [8]

2024 , eprint =

Enugandla, Shreepranav Varma and Fang, Xin and Fourier, Ghislain and Steinert, Christian , title =. 2024 , eprint =

work page 2024