REVIEW
Mutation Sequences along Weaves and Amalgamation of Braid Varieties
Algebraic Geometry · Submitted 5 Sep 2026 · 20 references
Pith's verdict: A solid structural result with a clear gap: the inductive sweep-mutation engine delegates the B2 and G2 checks to unperformed computations, so the main theorem is conditional on those checks holding up.
T0 · AI review
Why Pith says this
Highlight any text to comment on it.
Referee report
The paper
The paper claims
A quasi-cluster isomorphism links double Bott-Samelson strata to products of braid varieties.
The reading
What carries the argument
The sweep mutation sequence along a double Demazure weave: an inductively defined sequence of quiver mutations that tracks how the quiver changes as one reduces the double word via braid moves and contractions. At each contraction, a single mutation is performed at the middle string vertex, and that vertex is relabeled as a new frozen vertex (a trivalent vertex of the weave). After processing the entire weave, the quiver becomes an amalgamation of the terminal double-word quiver Q(δ) and the weave quiver Q(W), with the trivalent vertices providing the bridge.
Extended reading notes
Core claim
Theorem 1.1 (Corollary 5.24 and Theorem 5.25) proves that for positive braids p,q with Demazure products u,v, there exists a seed Σ on Conf(p,q) and a subset S of mutable vertices such that (1) the frozen seed describes the coordinate ring of Conf(p,q)_{u,v}, (2) the freezing is a cluster localization, and (3) the splitting isomorphism Conf(p,q)_{u,v} → Conf(u,v) × X(p) × X($q^{{op}}$) is a quasi-cluster isomorphism. The seed is built from a sweep mutation sequence along a double Demazure weave, one mutation per trivalent vertex, ending at an amalgamation of an extension of the double word quiver with the weave quiver.
Load-bearing premise
The inductive step of the sweep mutation sequence (Lemma 4.25) relies on verifying the exchange-matrix identity (55) for types B2 and G2 by finite computation 'left to the reader'; if those computations contain an error, the main mutation sequence may not terminate at the claimed quiver. Additionally, the trivalent identity of [CGGLSS25, Definition5.8, Lemma5.9] is imported without reproof.
Editorial extensions
If this is right
- The cluster structure on Conf(p,q)_{u,v} is explicitly described as a cluster localization of the known cluster structure on Conf(p,q).
- The splitting map provides a quasi-cluster isomorphism between the stratum and the product Conf(u,v) × X(p) × X(q^{op}), showing that cluster variables on the three factors map to cluster variables on the stratum.
- In the half-decorated setting, the splicing maps Ψ_{r,w_0} and Ψ_{N,e} of Gorsky–Kim–Scroggin–Simental are proved to be quasi-cluster isomorphisms, confirming their conjecture.
- The sweep mutation sequence provides an explicit combinatorial recipe to compute the cluster variables on the stratum from those on the whole cell.
Reading between the lines
- The sweep mutation sequence construction likely generalizes to other truncations of double Bott-Samelson cells beyond endpoint strata, providing cluster localizations for many open subvarieities.
- The explicit mutation sequence could be used to compute cluster variables for braid varieties from known formulas on double Bott-Samelson cells, offering a new computational tool.
- The relation to local systems on a tetrahedron suggests that sweep mutation sequences along weaves of type ΔΔ → Δ realize flips of ideal triangulations in higher Teichmüller theory for all Lie types.
Original abstract
Let $p,q$ be positive braids with Demazure products $u,v$. The endpoint stratum $\mathrm{Conf}(p,q)_{u,v}$ of the double Bott-Samelson cell splits as $\mathrm{Conf}(u,v) \times X(p) \times X(q^{\mathrm{op}})$, a double Bruhat cell times two braid varieties. We prove that the stratum's cluster structure is given by a cluster localization of the one on $\mathrm{Conf}(p,q)$, and that the splitting map is a quasi-cluster isomorphism. This comes from a mutation sequence along a double Demazure weave, one mutation per trivalent vertex, ending at an amalgamation of an extension of the double word quiver with the weave quiver. In the half-decorated case, this proves the conjecture of Gorsky-Kim-Scroggin-Simental for the splicing map $X(p) \times X(\Delta\Delta) \to X(p\Delta)$, where $\Delta$ is a reduced positive braid for $w_0$ and $\mathrm{dem}(p) = w_0$.
The compiled record
The paper's own sentences, arranged, kept beside the review and separate from it. Each label is drawn from a closed vocabulary; the number beside it is the probability the classifier gave that label. Numbers in the tables are the authors', checked by code. No sentence in this record was written by a model.
The abstract, sentence by sentence
Let $p,q$ be positive braids with Demazure products $u,v$. The endpoint stratum $\mathrm{Conf}(p,q)_{u,v}$ of the double Bott-Samelson cell splits as $\mathrm{Conf}(u,v) \times X(p) \times X(q^{\mathrm{op}})$, a double Bruhat cell times two braid varieties. We prove that the stratum's cluster structure is given by a cluster localization of the one on $\mathrm{Conf}(p,q)$, and that the splitting map is a quasi-cluster isomorphism. This comes from a mutation sequence along a double Demazure weave, one mutation per trivalent vertex, ending at an amalgamation of an extension of the double word quiver with the weave quiver. In the half-decorated case, this proves the conjecture of Gorsky-Kim-Scroggin-Simental for the splicing map $X(p) \times X(\Delta\Delta) \to X(p\Delta)$, where $\Delta$ is a reduced positive braid for $w_0$ and $\mathrm{dem}(p) = w_0$.
background contribution method result interpretation main claim
| # | sentence | role | the authors' own work | states the main conclusion | strength |
|---|---|---|---|---|---|
| s0 | Let $p,q$ be positive braids with Demazure products $u,v$. | background 1.00 | 0.07 | 0.02 | unclear 0.65 |
| s1 | The endpoint stratum $\mathrm{Conf}(p,q)_{u,v}$ of the double Bott-Samelson cell splits as $\mathrm{Conf}(u,v) \times X(p) \times X(q^{\mathrm{op}})$, a double Bruhat cell times two braid varieties. | result 0.91 | 0.44 | 0.17 | asserted 1.00 |
| s2 | We prove that the stratum's cluster structure is given by a cluster localization of the one on $\mathrm{Conf}(p,q)$, and that the splitting map is a quasi-cluster isomorphism. | result 0.90 | 0.95 | 0.57 | asserted 1.00 |
| s3 | This comes from a mutation sequence along a double Demazure weave, one mutation per trivalent vertex, ending at an amalgamation of an extension of the double word quiver with the weave quiver. | method 0.96 | 0.87 | 0.21 | asserted 0.99 |
| s4 | In the half-decorated case, this proves the conjecture of Gorsky-Kim-Scroggin-Simental for the splicing map $X(p) \times X(\Delta\Delta) \to X(p\Delta)$, where $\Delta$ is a reduced positive braid for $w_0$ and $\mathrm{dem}(p) = w_0$. | result 0.50 | 0.84 | 0.18 | asserted 1.00 |
What was examined, and what was not
- arXiv listing abstract v1
- full text
- figures and tables
- references and citing sentences
- relations to other papers
- P abstract
- S sections
- T body sentences
- C citations
- R cross-paper relations
Figures
Figures reproduced from arXiv:2609.05833 by the authors.
References
Discussion
Disagree with the review, or know something it missed? Leave a comment, highlight any passage above to comment on it, or challenge this review formally.