Pith. sign in

REVIEW 3 major objections 5 minor 21 references

Hybrid pipe dreams for the lower-upper scheme

T0 review · 3 major / 5 minor · reviewed 2026-08-05 · deepseek-v4-flash

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

Pith's one-line read Generic pipe dream polynomials are hybridization-independent and compute lower-upper equivariant classes.

desk verdict Strong paper with a hinge that needs to be made visible: the d=1 CAS check in Theorem 21. read the letter →

arxiv 2509.01857 v1 pith:UWAYCORW submitted 2025-09-02 math.CO math.AG

Allen Knutson , Paul Zinn-Justin This is my paper
classification math.COmath.AG MSC 14M1505E0514N1582B23
keywords genericpipedreamshybridSchubertpolynomialsequivariantcohomologylower-uppervarietiesYang-Baxterequationcompleteintersectionsfluxvariables
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

This paper aims to show that the classic and bumpless pipe-dream models for Schubert polynomials are not merely equivalent counting devices, but are both visible inside one geometric object. It introduces hybrid generic pipe dreams, where each row independently chooses a West or East convention, and proves that the resulting polynomial Gπ is independent of that choice. It then identifies Gπ with (A+B)^m times the torus-equivariant cohomology class of a lower-upper variety Eπ, giving the polynomial a geometric meaning. The main degeneration construction breaks Eπ into a union of quadratic complete intersections indexed by hybrid generic pipe dreams of connectivity π, so individual pipe dreams become components of an explicit degeneration.

What carries the argument

The named object is the hybrid generic pipe dream polynomial Gπ: a weighted state sum over fillings of an m×n grid by elbow, straight, and blank tiles, each row independently typed W or E, with weights linear in variables A, B, xi, yj. The carrying identities are the Yang–Baxter equations for the tile weights, which force β-independence and the divided-difference recurrence, and the flux variables attached to the edges of the lower-upper scheme. Flux conservation at each square is what reconstructs a hybrid generic pipe dream from the components of the degeneration, turning the drawings into geometry.

What would settle it

Recompute the ideal I = ⟨(gSg^{-1})_{12} = 0, (gSg^{-1})_{11} = a, (gSg^{-1})_{22} = e⟩ symbolically with generic a,e, and check whether the q ≠ 0 locus is empty after discarding the three exceptional components a = e = 0, a = e, and a = -e. A single point with q ≠ 0 satisfying the equations would break the d = 1 step and hence Theorem 21. A second independent check would be to compute Gπ by the recurrence of Theorem 17 and by the degeneration of Theorem 30 for a small example such as m = n = 3, π = 312, and compare the two polynomials termwise.

Watch

Extended reading notes

Core claim

The central claim is Gπ = (A+B)^m [Eπ], with Gπ the generic pipe dream polynomial and [Eπ] the torus-equivariant cohomology class of the lower-upper variety Eπ (pairs of matrices whose products are triangular, restricted to component π). Yang–Baxter equations prove Gπ independent of the West/East hybridization and give a divided-difference recurrence. Two geometric proofs follow: one checks the recurrence on the classes, the other degenerates Eπ into a union Vδ of quadratic complete intersections indexed by hybrid generic pipe dreams of connectivity π, each Vδ contributing exactly one summand of Gπ. The B → ∞ limit recovers the double Schubert polynomial, with nongeneric hybrid pipe dreams s

Load-bearing premise

The proof that the equivariant classes satisfy the divided-difference recurrence depends on an unshown computer algebra calculation that a certain counting degree d is 1 for the 2-by-2 matrix group action on Eπ; if that calculation were wrong for some π, the recurrence—and the equality with the pipe-dream polynomial—would carry an extra factor.

Editorial extensions

If this is right

  • If Theorem 2 and Corollary 22 hold, Gπ is a well-defined invariant of π computed by any row-by-row hybridization; classic and bumpless pipe-dream formulas become limiting cases of one polynomial.
  • Theorem 17 gives an inductive definition of Gπ with an explicit base case, so equivariant classes can be computed by a divided-difference recurrence without enumerating pipe dreams.
  • Theorem 24 shows the double Schubert polynomial is the B-leading form of Gπ, geometrically tying Schubert polynomials to the lower-upper scheme.
  • The degeneration into complete intersections means each pipe-dream component Vδ carries an explicit equivariant class and is generically reduced; if the conjectured absence of embedded components holds, the geometric decomposition exactly matches the pipe-dream summands.
  • Because the construction is inductive on m and works for rectangular m ≤ n, the lower-upper scheme can be handled without restricting to square matrices, which the proof needs for the induction step.

Reading between the lines

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

  • The paper does not pursue a term-by-term bijection between hybridizations, and the counts 76, 78, 80 for π=1253 show no naive bijection exists; a natural extension would be to construct the decorated bijection suggested by the 1475 decorated generic pipe dreams.
  • The flux-equation reformulation suggests a definition of pipe dream that depends only on flux equalities and boundary data rather than declared tile shapes; if developed, this could produce analogues for other boundary conditions or other matrix varieties.
  • Because the degeneration is only partial—weights may tie rather than form a full monomial order—the natural geometric pieces are nonlinear complete intersections rather than coordinate subspaces; a further degeneration might recover the nonreduced phenomena seen in the bumpless pipe-dream Gröbner geometry.
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.

Referee Report

3 major / 5 minor

Summary. The paper introduces hybrid generic pipe dreams for rectangular lower-upper schemes. It proves, via Yang–Baxter state-sum identities, that the associated polynomial Gπ is independent of the hybridization β, and that it satisfies a divided-difference recurrence. It then identifies Gπ with (A+B)^m times the torus-equivariant cohomology class of the lower-upper variety Eπ, and gives a degeneration of Eπ into complete intersections indexed by the hybrid generic pipe dreams, with class contributions matching the terms of Gπ. A flux formalism is introduced that motivates the pipe-dream combinatorics from the degenerating equations.

Significance. If correct, the paper gives a geometric realization of hybrid generic pipe dreams: each pipe dream corresponds to a component of a flat degeneration, a substantial strengthening of the formal role pipe dreams usually play. The equivariant setting and the Yang–Baxter proof of hybridization independence are new relative to [KU23], and the flux reinterpretation is an appealing conceptual contribution. The main objects are independently defined, no fitted parameters appear, and the paper contains a number of checkable recurrence and degeneration statements. However, several load-bearing verifications are abbreviated, so the result is plausible but not yet fully verified as written.

major comments (3)
  1. [§4.1, proof of Theorem 21, Eq. (4.2)] The recurrence (4.2) is the geometric input to Corollary 22 and hence to the central identity Gπ=(A+B)^m[Eπ]. The proof that d=1 in Lemma 20 is delegated to an unseen computer algebra computation: the ideal I is declared to have four components, but the component analysis is not displayed. The displayed ideal also contains a notation error: S=diag(a,e), g=[[p,q],[r,s]], but the third generator is written as (gSg^{-1})_{22}=d, and the entry s clashes with the matrix S. A wrong d would introduce a factor 1/d in the divided difference operator and break the comparison with Theorem 17. This verification must be supplied, either as a complete computer-algebra transcript or as a hand proof, with the notation corrected.
  2. [§3.1, Propositions 11 and 12] Theorem 16 (independence of β) and Theorem 17 (divided-difference recurrence) rest on the two Yang–Baxter state-sum identities. The proofs state 'direct computation' and illustrate only two connectivity cases for Proposition 11; Proposition 12 is dismissed by symmetry. Since the state sums involve several possible connectivities, the reader cannot verify the claimed identities from the text. Please list the finite set of connectivity patterns that must be checked and include the checks, or give a precise reference where the full computation is carried out.
  3. [§5.4 and Theorem 30] The degeneration theorem is stated with 'possibly ... embedded components,' but the class comparison in (5.14)–(5.15) yields an equality [Eπ] = Σ [Vδ]. In the positive multigrading of §4.2, nonzero embedded components would contribute additional classes, so the displayed inequality argument cannot by itself establish both the class equality and the presence of embedded components. The paper should either prove that the class equality rules out embedded components and remove the caveat from Theorems 2 and 30, or restate the degeneration result as a set-theoretic degeneration plus a separate class computation. As written, the claim that pipe dreams are exactly the components of the degeneration is not fully resolved.
minor comments (5)
  1. [§4.1, Theorem 21 proof] The typo in the ideal I — (gSg^{-1})_{22}=d should involve e, not d — should be corrected; also the matrix entry r/s notation should be changed to avoid confusion with the diagonal matrix S.
  2. [§3.1, Lemma 13] The sentence 'This correspondence also appears as part of a larger bijection in the very recent [Wei25], but we cannot see any connection' is informal and not needed; either explain the relevance or delete it.
  3. [§1.3.3, example for m=n=2, π=12] The example refers to 'underlined' terms, but underlining is not visible in the submitted text. Please mark the retained terms explicitly, e.g., by boldface.
  4. [§5.6, Proposition 31] The proof is a sketch ending with 'We leave it as an exercise to the reader.' Since Proposition 31 is a stated geometric identification, either give the full verification or label the statement as a sketch/proof omitted.
  5. [§1.3.2, Theorem 2] The sentence 'Part (2) follows immediately from part (1)' is slightly misleading because part (1) is qualified by possible embedded components; the independent proof of part (2) in §4 is welcome, but the wording should be adjusted.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found: Gπ and [Eπ] are independently defined and matched through their own recurrence proofs; the one CAS verification gap is a non-circular rigor issue.

full rationale

The paper's central identity Gπ=(A+B)^m[Eπ] (Cor. 22) is obtained by comparing two independently established recurrence structures. Gπ is defined combinatorially as a state sum over hybrid generic pipe dreams (§1.2) and shown by Yang–Baxter computations (Props. 11–12, Lemma 13) to be β-independent (Thm. 16) and to satisfy the divided-difference recurrence (3.2) (Thm. 17). The equivariant class [Eπ] of the lower-upper variety is defined geometrically (§1.3.1, Props. 1, 6) and proved, via the geometric divided-difference lemma (Lemma 20, from [BBM89]) and a GL2-action analysis, to satisfy the same recurrence (4.2) and base case (4.1) (Thm. 21). The two proofs do not assume the equality being proved. The degeneration theorem (Thm. 30) constructs varieties Vδ whose equations (including flux equations) are derived from the geometry, and computes [Vδ] using the already-proved Cor. 22; the class comparison in §5.4 turns an inclusion of schemes into equality using the positivity of the multigrading (§4.2), not by assuming the target formula. Self-citations to [Knu05], [KU23], [KZJ24] are disclosed and used for context or announced results; the load-bearing lemma 20 is from [BBM89] and [KZJ14, Prop. 3] is a general proposition whose assumptions do not include the target identity. The only flagged gap is the computer-algebra check in Thm. 21 that the degree d in Lemma 20 equals 1: the manuscript states "This is then a computer algebra calculation ... the ideal I = ⟨(gSg^{-1})_{12}=0, (gSg^{-1})_{11}=a, (gSg^{-1})_{22}=d⟩ is computed to have four components ..." without displaying the computation or code, and with an apparent typo (d versus e). This is an omitted verification of a numerical input, not a self-referential definition or a fitted parameter; it affects rigor, but the derivation chain is not circular. If the computation were wrong, Cor. 22 would fail, but the failure would be an unsupported hinge, not a reduction of the conclusion to the premise. Therefore no circularity is found.

Assumptions & free parameters 0 free parameters · 4 assumptions · 2 invented entities

The paper is self-contained for the combinatorial state sums; the geometric part relies on standard equivariant cohomology, a divided-difference lemma from [BBM89], and the B-leading form proposition from [KZJ14]. No numerical fitting or hidden parameters enter.

assumptions (4)
  • standard math Lemma 20 of [BBM89] (geometric divided differences for GL2 actions)
    Quoted in full and used to convert the geometry of GL2 saturation into the divided difference recurrence for [Eπ] in the proof of Theorem 21.
  • standard math Proposition 3 of [KZJ14] (B-leading forms under projection)
    Used in Theorem 24 to relate the B-leading form of [Eπ] to the Schubert polynomial of the matrix Schubert variety under the projection (X,Y) -> X.
  • standard math Equivariant cohomology identification H_T^*(V) ≅ H_T^*(pt) and positivity of multigrading ([AF24], [MS05])
    Background: identifies classes with polynomials and justifies the inequality [X] ≤ [Y] used in the equality argument of §5.4.
  • standard math Flatness of weight degenerations from C* actions (Rees algebra / associated graded)
    Assumed in §5 when calling the partial Gröbner limit a degeneration; standard but not explicitly proven in the text.
invented entities (2)
  • Flux variables Φ_V(i,j), Φ_H(i,j) independent evidence
    purpose: Edge labels built from X_{ij}Y_{ji}; their conservation and equalities determine the pipe dream attached to a degeneration component.
    Defined from the matrix entries rather than postulated; the paper proves flux conservation and derives the pipe dream from flux equalities (Prop. 3, Theorem 30).
  • Hybrid generic pipe dreams independent evidence
    purpose: Index the terms of Gπ and the components of the degeneration of Eπ.
    Introduced with explicit tile weights; their state sums are shown to be independent of β and to match geometric classes, providing internal evidence.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Hybrid pipe dreams for the lower-upper scheme." pith.science (2026). https://pith.science/paper/UWAYCORW

@misc{pith2026250901857,
  author       = {Pith},
  title        = {Pith review of: Hybrid pipe dreams for the lower-upper scheme},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/UWAYCORW}},
  note         = {Machine review of arXiv:2509.01857}
}
read the original abstract

In [KU23] were introduced hybrid pipe dreams interpolating between classic and bumpless pipe dreams, each hybridization giving a different formula for double Schubert polynomials. A bijective proof was given (following [GH23]) of the independence of hybridization, but only for nonequivariant Schubert polynomials. In this paper we further generalize to hybrid generic pipe dreams, replacing the bijective proof of hybridization-independence with a Yang-Baxter-based proof that allows one to maintain equivariance. An additional YB-based proof establishes a divided-difference type recurrence for these generic pipe dream polynomials. These polynomials compute something richer than double Schubert polynomials, namely the equivariant classes of the lower-upper varieties introduced in [Knu05]. We give two proofs of this: the easier being a proof that the recurrence relation holds on those classes, the more difficult being a degeneration of the lower-upper variety to a union of quadratic complete intersections (plus, possibly, some embedded components) whose individual classes match those of the generic pipe dreams. One new feature of the generic situation is a definition of the "flux" through an edge of the matrix; the notion of pipe dream itself can then be derived from the equalities among the fluxes.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

21 extracted references · 17 canonical work pages

  1. [1]

    David Anderson and William Fulton, Equivariant cohomology in algebraic geometry, Cambridge University Press, 2024, https://people.math.osu.edu/anderson.2804/ecag/index.html

    work page 2024

  2. [2]

    4, 257 -- 269

    Nantel Bergeron and Sara Billey, RC -graphs and S chubert polynomials , Experimental Mathematics 2 (1993), no. 4, 257 -- 269

    work page 1993

  3. [3]

    Borho, J.-L

    W. Borho, J.-L. Brylinski, and R. MacPherson, Nilpotent orbits, primitive ideals, and characteristic classes. a geometric perspective in ring theory, Progress in Mathematics, vol. 78, Birkh \"a user Boston Inc., Boston, MA, 1989

    work page 1989

  4. [4]

    Philippe Di Francesco and Paul Zinn-Justin, Inhomogeneous model of crossing loops and multidegrees of some algebraic varieties, Comm. Math. Phys. 262 (2006), no. 2, 459--487, http://arxiv.org/abs/math-ph/0412031 arXiv:math-ph/0412031 . MR2200268 (2006k:82068)

    work page Pith review arXiv 2006

  5. [5]

    Sergey Fomin and Richard Stanley, Schubert polynomials and the nil- C oxeter algebra , Adv. Math. 103 (1994), no. 2, 196--207. MR1265793 (95f:05115)

    work page 1994

  6. [6]

    William Fulton, Flags, S chubert polynomials, degeneracy loci, and determinantal formulas , Duke Math. J. 65 (1992), no. 3, 381--420. 1154177

    1992

  7. [7]

    21, 18629--18663

    Yibo Gao and Daoji Huang, The canonical bijection between pipe dreams and bumpless pipe dreams, International Mathematics Research Notices 2023 (2023), no. 21, 18629--18663

    work page 2023

  8. [8]

    Daoji Huang, Bijective proofs of M onk's rule for S chubert and double S chubert polynomials with bumpless pipe dreams , Electron. J. Combin. 30 (2023), P3.4, http://arxiv.org/abs/2010.15048 arXiv:2010.15048

    work page Pith review arXiv 2023

Show all 21 references
  1. [9]

    Allen Knutson and Ezra Miller, Gr\"obner geometry of S chubert polynomials , Ann. of Math. (2) 161 (2005), no. 3, 1245--1318. MR2180402 (2006i:05177)

    2005

  2. [10]

    Algebraic Geom

    Allen Knutson, Some schemes related to the commuting variety, J. Algebraic Geom. 14 (2005), no. 2, 283--294, http://arxiv.org/abs/math.AG/0306275 arXiv:math.AG/0306275 . MR2123231 (2005j:14076)

    2005

  3. [11]

    63–83, Cambridge University Press, 2022, http://arxiv.org/abs/1909.13777 arXiv:1909.13777

    , Schubert polynomials, pipe dreams, equivariant classes, and a co-transition formula, London Mathematical Society Lecture Note Series, p. 63–83, Cambridge University Press, 2022, http://arxiv.org/abs/1909.13777 arXiv:1909.13777

    2022 arXiv

  4. [12]

    Allen Knutson and Gabe Udell, Interpolating between classic and bumpless pipe dreams, S\'em. Lothar. Combin. 89B (2023)

    2023

  5. [13]

    Patricia Klein and Anna Weigandt, Bumpless pipe dreams encode G r\"obner geometry of S chubert polynomials , 2021, http://arxiv.org/abs/2108.08370 arXiv:2108.08370

    2021 arXiv

  6. [14]

    Allen Knutson and Paul Zinn-Justin, A scheme related to the B rauer loop model , Adv. Math. 214 (2007), no. 1, 40--77, http://arxiv.org/abs/math.AG/0503224 arXiv:math.AG/0503224 . MR2348022 (2008j:14097)

    2007

  7. [15]

    , The B rauer loop scheme and orbital varieties , J. Geom. Phys. 78 (2014), 80--110, http://arxiv.org/abs/1001.3335 arXiv:1001.3335

    2014 arXiv

  8. [16]

    , Generic pipe dreams, lower-upper varieties, and S chwartz-- M ac P herson classes , 2024, http://arxiv.org/abs/2411.11208 arXiv:2411.11208

    2024 arXiv

  9. [17]

    5, 883–962, http://arxiv.org/abs/1806.11233 arXiv:1806.11233

    Thomas Lam, Seung Jin Lee, and Mark Shimozono, Back stable S chubert calculus , Compositio Mathematica 157 (2021), no. 5, 883–962, http://arxiv.org/abs/1806.11233 arXiv:1806.11233

    2021 arXiv

  10. [18]

    227, Springer-Verlag, New York, 2005

    Ezra Miller and Bernd Sturmfels, Combinatorial commutative algebra, Graduate Texts in Mathematics, vol. 227, Springer-Verlag, New York, 2005. 2110098 (2006d:13001)

    2005

  11. [19]

    Brian Rothbach, Equidimensionality of the B rauer loop scheme , Electron. J. Combin. 17 (2010), no. 1, Research Paper 75, 11, http://www.combinatorics.org/Volume_17/Abstracts/v17i1r75.html. 2651728

    2010

  12. [20]

    Anna Weigandt, Bumpless pipe dreams and alternating sign matrices, Journal of Combinatorial Theory, Series A 182 (2021), 105470, http://arxiv.org/abs/2003.07342 arXiv:2003.07342

    2021 arXiv

  13. [21]

    , Changing bases with pipe dream combinatorics, 2025, http://arxiv.org/abs/2506.07306 arXiv:2506.07306

    2025 arXiv

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.