Pith. sign in

REVIEW 5 major objections 5 minor 7 references

The Last Three T-degrees in Triply-Graded Link Homology

T0 review · 5 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read For any braid, the outermost three T-degree layers of reduced triply-graded HOMFLY homology are determined explicitly, and are almost entirely zero with only a few one-dimensional k-modules.

desk verdict Computes the extreme T-degree part of HHH for all positive/negative braid closures as R-modules; the core computations look plausible, but the 'all braids' claim rests on a sketchy braid-word lemma and a few compressed proofs. read the letter →

arxiv 2505.14182 v1 pith:BOWV7KB6 submitted 2025-05-20 math.RT math.GTmath.QA

Cailan Li This is my paper · ORCID
classification math.RTmath.GTmath.QA
keywords homologydegreeslinkmathrmnegativeoverlinethreeapproach
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

Knots and links can be represented by braids, where strands cross over and under each other. Triply-graded link homology is a sophisticated invariant that takes a link and produces a graded vector space, refining the HOMFLY polynomial. Computing it is hard because it is the homology of a homology: one first builds a complex from a braid, then applies Hochschild cohomology, then takes cohomology again.

This paper studies only the extreme homological degrees: the three highest degrees for positive braids (all crossings positive) and the three lowest for negative braids. The main finding is that these extreme pieces are almost always zero. For a positive braid on the closure, the very top degree contains exactly a one-dimensional piece in one of the three internal gradings; the next degree is zero; and the third-from-top contains one or two one-dimensional pieces depending on the number of strands. For negative braids on three strands, there is a single non-zero piece in the third-from-bottom degree; for four or more strands, even that vanishes.

The computations are done by hand using a diagrammatic calculus for Hochschild cohomology of Soergel bimodules, a tool developed in earlier work by the same author. The uniformity is striking: the answer in these extreme degrees depends very little on the particular braid word. The paper also derives consequences for the HOMFLY polynomial of braid closures, giving necessary conditions that a link must satisfy to be a positive or negative braid link.

Extended reading notes

Core claim

Main Theorem: "We compute HHH^{A,T,Q}(β) as an R-module in the three highest (lowest) T-degrees for any positive (negative) braid β ∈ B_n for all values of A,Q." Concretely, for positive braids, T=|β| is k(|β|) at A=0 and zero otherwise; T=|β|-1 vanishes; T=|β|-2 is k(|β|-4) at A=0, k(|β|) at A=1, and zero for A≥2, under connectivity conditions. For negative braids on n≥3, the two lowest degrees vanish for all A; at degree -|α|+2, only 3-strand braids have a nonzero class, k(8-|α|) at A=2.

Load-bearing premise

Proposition 2.4: a braid β with minimal strand count n for its closure bβ must contain σ_iσ_{i+1}σ_iσ_{i+1} or σ_{i+1}σ_iσ_{i+1}σ_i as a subexpression for every 1 ≤ i ≤ n−2. This is what lets the Main Theorem claim to cover all positive and negative braids, rather than only those whose words contain the special alternating patterns. The proof is a braid-move argument that is only sketched (it analyzes the 'final move' after a sequence of braid relations); if a minimal braid word avoided these subexpressions, Theorems C and D would apply only to a restricted subclass.

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

5 major / 5 minor

Summary. The paper computes the reduced triply-graded link homology HHH as an R-module in the three top (resp. bottom) T-degrees for closures of arbitrary positive (resp. negative) braids. For positive braids, it claims that T=|β| is concentrated in A=0 degree k(|β|), that T=|β|-1 vanishes, and that T=|β|-2 is k(|β|-4) at A=0, k(|β|) at A=1, and zero for A≥2, under a connectivity condition. For negative braids on n≥3 strands, the two lowest degrees vanish; for 3-strand negative braids there is a nontrivial class at T=-|α|+2, A=2. The proofs use diagrammatic Hochschild cohomology of Soergel bimodules and a reduction (Proposition 2.4) to braids containing alternating length-4 subexpressions.

Significance. If the computations are correct, the paper establishes a striking uniformity in the top T-degrees of triply-graded link homology and gives the first such computations for arbitrary braids, with applications to necessary conditions for positive/negative braid links and to a possible spectral sequence to the (2,3)-torus knot. The explicit kernel/image matrices in Theorem 4.1 and the emphasis on R-module structure (rather than only vector-space grading) are concrete strengths, and the main algebraic results are derived by direct computation rather than by fitting parameters. However, several reduction and base-case proofs are too compressed to verify the full scope of the claims, so the main theorem is plausible but not fully established in the current text.

major comments (5)
  1. [Proposition 2.4] The proof of the reverse implication is not sufficient to justify that every positive braid of minimal strand count contains σ_iσ_{i+1}σ_iσ_{i+1} or σ_{i+1}σ_iσ_{i+1}σ_i as a subexpression for every i. The argument analyzes only the 'final move' (A)/(B) in a sequence of braid moves and asserts without proof that the surrounding letters contain no s or t; it does not rule out final moves that involve several adjacent pairs at once, nor does it prove that a violation for a fixed k0 can be introduced only by moves (A)/(B). The minimality argument involving Markov move II also needs a more detailed justification. Since Proposition 2.4 is what allows Theorems C, D, and the Main Theorem to apply to all positive/negative braids rather than only to those already containing the special subexpressions, a complete proof is load-bearing.
  2. [Lemma 2.3] The case analysis is incomplete: the proof explicitly treats the sstt subexpression and then asserts that the cases stts and tsst can be reduced to sstt by Markov move I, but a cyclic permutation changes the cyclic word and does not obviously preserve the non-connect-sum hypothesis while converting the subexpression. Since Lemma 2.3 feeds directly into Proposition 2.4, this gap propagates to the claimed scope of the Main Theorem.
  3. [Section 2.5] The formulas for HHH(σ_1^m) and HHH(σ_1^{-m}) in Section 2.5 are stated as facts without proof or reference. These formulas are the n=2 base case of the Main Theorem and are also used in the applications in Theorem 1.2. Without a derivation or a citation, the claim that the Main Theorem covers the n=2 case is unsupported.
  4. [Theorem 4.2] The proof is compressed to the assertions that 'the kernels are fairly easy to work out' and that the only nonzero cohomology is generated by a single displayed element. Since this theorem is one of the main results (the negative 3-strand case), the reader cannot verify the cancellation of all other cohomology groups from the material given. A complete kernel/image computation, or at least a detailed indication of how Lemma 3.4 dualizes the explicit matrices of Theorem 4.1, is needed.
  5. [Theorem D] The proof outlines the structure of the relevant matrices but relies on 'as a result' and 'it is then clear that this is the image of HH^k(d_{|β|-2})^T' after a single illustrative example. The definitions of N, ℓ, and the block decomposition are not fully justified, and the final identification of the kernel of the transposed differential with the image of HH^k(d_{|β|-2})^T is asserted rather than shown. Given that Theorem D is a central component of the Main Theorem for n≥4, this proof needs to be written out in more detail.
minor comments (5)
  1. [Theorem C and Theorem D] Both statements quantify 'for all 1 ≤ i ≤ n - 1' where σ_iσ_{i+1} appears; this should be '1 ≤ i ≤ n - 2', since σ_{i+1} is undefined for i = n-1.
  2. [Section 2.3] The distinction between 'subword' and 'subexpression' is confusing; the example says 'there is only one subword of length 1 of ss while there are 2 possible subexpressions', but the earlier definition of subexpression allows repeated letters. These terms should be defined explicitly and used consistently.
  3. [Section 4.1, Theorem 4.2] The displayed generator for the nonzero cohomology class in Theorem 4.2 is difficult to parse in text form; a diagram or an explicit vector in the basis from Appendix A would improve the presentation.
  4. [Section 1.2.1, proof of Theorem 1.2(2)] The sentence 'Note such a braid representative also exists for T(2,k) as σ_1^k ∼ σ_1^k σ_2 ∼ ...' is not fully argued; the chain of equivalences is unclear and should be expanded or justified.
  5. [Abstract and Section 1] The notation alternates between \overline{\mathrm{HHH}} (abstract) and HHH (body). Please standardize the notation.
Assumptions & free parameters 0 free parameters · 7 assumptions · 0 invented entities

The paper is a computational tour de force that introduces no free parameters or new entities. Its load-bearing inputs are the Soergel-bimodule framework, the author's own diagrammatic Ext calculus [Li22], Serre duality, and the braid-reduction classification. The only unproven internal input is the recorded T(2,±k) base case, and the black-box use of the diagrammatic basis is explicitly flagged by the author.

assumptions (7)
  • domain assumption Rouquier complex F_β and Hochschild cohomology HH^k(-) produce HHH with three gradings (A,T,Q).
    Section 1.1 takes Khovanov's construction [Kho07, Rou04] as the definition of the invariant.
  • domain assumption Diagrammatic bases for HH^k(B_w) for w a word in Soergel bimodules, including B_sB_tB_s and B_sB_tB_u, as in Appendix A and [Li22].
    All matrix computations in Sections 3 to 5 rest on these explicit bases; the paper says 'take Appendix A as a black box'.
  • domain assumption Serre duality HH^k(F_{β∨}) ≅ Hom^•_R(HH^{n-1-k}(F_β), R)(2(n-1)) (Lemma 3.4, from [GHMN19]).
    Used to transfer all positive-braid results to negative braids; a grading or degree error would shift the negative-braid vanishing statements.
  • standard math Koszul complex exactness when any parameter is a unit (Corollary 2.7).
    Used in Theorem A and Theorem 5.6 to kill cohomology in the J-summands.
  • domain assumption Connect sum formula HHH(β1#β2) = HHH(β1) ⊗_k HHH(β2) (Proposition 2.5, cited from [Ras15]).
    Allows reduction to closures that are not connect sums.
  • domain assumption Braid-index reduction: any positive braid with minimal strand count n contains σ_iσ_{i+1}σ_iσ_{i+1} or reverse subexpressions for all i (Proposition 2.4).
    The proof is sketched with a braid-move argument; this premise extends Theorems C and D from the special braids to all positive and negative braids.
  • domain assumption R-module structure of HHH(σ_1^{±m}) as recorded in Section 2.5.
    Used as base case for n=2; no proof or citation is given in the text.

how reviews work

0 comments
Cite this review

Pith. "Pith review of The Last Three T-degrees in Triply-Graded Link Homology." pith.science (2026). https://pith.science/paper/BOWV7KB6

@misc{pith2026250514182,
  author       = {Pith},
  title        = {Pith review of: The Last Three T-degrees in Triply-Graded Link Homology},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/BOWV7KB6}},
  note         = {Machine review of arXiv:2505.14182}
}
abstract

We investigate the structure of reduced triply graded link homology $\overline{\mathrm{HHH}}$ in the top/bottom three $T-$degrees for links arising as closures of positive/negative braids. Using a diagrammatic approach to the Hochschild cohomology of Soergel bimodules, we provide explicit computations of $\overline{\mathrm{HHH}}$ as $R-$modules in these degrees. Our results reveal that the homology here is often zero, especially in the negative braid case, and display striking uniformity.

Figures

Figures reproduced from arXiv: 2505.14182 by the authors.

Figure 1
Figure 1. The green area represents the results of [ [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

7 extracted references · 5 canonical work pages

  1. [1]

    Koszul duality for Kac-Moody groups and characters of tilting modules

    [AMRW19] P. N. Achar, S. Makisumi, S. Riche, and G. Williamson. “Koszul duality for Kac-Moody groups and characters of tilting modules”. In:J. Amer. Math. Soc.32.1 (2019), pp. 261–310. [BCH23] C. Bowman, A. Cox, and A. Hazi. “Path isomorphisms between quiver Hecke and diagrammatic Bott- Samelson endomorphism algebras”. In:Adv. Math.429 (2023). [BM90] J.S....

    work page 2019

  2. [6]

    Reducing Rouquier complexes

    eprint:https: //makisumi.com/math/articles/hsbim.pdf. [Mal24] L. Maltoni. “Reducing Rouquier complexes”. In: Proc. Lond. Math. Soc. (3)129.1 (2024). [NS24] K.NakaganeandT.Sano. ComputationsofHOMFLYhomology .2024.arXiv: 2111.00388 [math.GT]. [Pic20] L. Piccirillo. “The Conway knot is not slice”. In: Ann. of Math. (2)191.2 (2020), pp. 581–591. [Ras10] J. Ra...

    work page arXiv 2024

  3. [58]

    The two-color Soergel calculus

    [Eli16] B. Elias. “The two-color Soergel calculus”. In: Compos. Math.152.2 (2016), pp. 327–398. [EQ23] B.EliasandY.Qi.“CategorifyingHeckealgebrasatprimerootsofunity,partI”.In: Trans.Amer.Math. Soc.376.11 (2023), pp. 7691–7742. [EW14] B. Elias and G. Williamson. “The Hodge theory of Soergel bimodules”. In: Ann. of Math. (2)180.3 (2014), pp. 1089–1136. [EW1...

    work page 2016

  4. [426]

    Triply-graded link homology and Hochschild homology of Soergel bimodules

    [Kho07] M. Khovanov. “Triply-graded link homology and Hochschild homology of Soergel bimodules”. In: Internat. J. Math.18.8 (2007), pp. 869–885. [Li22] C. Li. The Two-Color Ext Soergel Calculus

    work page 2007

  5. [2004]

    Kategorie O, perverse Garben und Moduln über den Koinvarianten zur Weylgruppe

    arXiv:math/0409593 [math.RT]. [Soe90] W. Soergel. “Kategorie O, perverse Garben und Moduln über den Koinvarianten zur Weylgruppe”. In: J. Amer. Math. Soc.3.2 (1990), pp. 421–445. [Wil17] G. Williamson. “Schubert calculus and torsion explosion”. In: J. Amer. Math. Soc.30.4 (2017). With a joint appendix with Alex Kontorovich and Peter J. McNamara, pp. 1023–1046. 22

    arXiv 1990

  6. [2019]

    Categorified Young symmetrizers and stable homology of torus links

    arXiv:1909.00418. [Hog18] M. Hogancamp. “Categorified Young symmetrizers and stable homology of torus links”. In: Geom. Topol.22.5 (2018), pp. 2943–3002. [Kho00] M. Khovanov. “A categorification of the Jones polynomial”. In: Duke Math. J.101.3 (2000), pp. 359–

    arXiv 2018

  7. [2022]

    The Two-Color Ext Soergel Calculus

    arXiv:2211.07802 [math.RT]. [LW22] N. Libedinsky and G. Williamson. “The anti-spherical category”. In: Adv. Math.405 (2022). [Mak22] S. Makisumi. Diagrammatics for Ext-Enhanced Soergel Bimodules in TypeA1

    work page Pith review arXiv 2022

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.