REVIEW 3 major objections 4 minor 17 references
Insensitivity of Khovanov homology under rim surgery
T0 review · 3 major / 4 minor · reviewed 2026-08-01 · deepseek-v4-flash
Pith's one-line read This paper proves that rim surgery, and any local annulus replacement, leaves the map induced by a surface on Khovanov homology unchanged over every coefficient ring.
desk verdict Answers Hayden-Sundberg's question with a likely-true theorem, but the first proof contains a false disk-existence claim and the full-strength statement depends on an unpublished thesis. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The load-bearing object is the Khovanov skein lasagna module, a 4-dimensional refinement of Khovanov homology that assigns bigraded groups to pairs (X,L) of a 4-manifold and a framed link in its boundary. The first proof uses the original 0-dimensional-input version over a field of characteristic not 2, and the second uses the 1-dimensional-input version whose input theory is Rozansky–Willis homology for links in connected sums of S^1×S^2. The decisive computation is that the relevant graded piece S^{2,1d,0-div}_{0,0,0}(S^1×B^3; 1_2) — where 1_2 is the standardly framed two-component link S^1×∂B^1 in S^1×S^2 — is isomorphic to R, free of rank one. A cap-off argument in S^4 then shows that ev
What would settle it
Over Z/2, take a genus-1 surface in B^4 bounding the unknot, perform a rim surgery along a null-homologous curve with the right-handed trefoil pattern using the even framing, and compute the two induced maps Kh(∅;Z/2) → Kh(L;Z/2). The paper predicts they are equal; any difference is a counterexample to Theorem 1. Alternatively, a gap in the unpublished functoriality proof for Rozansky–Willis homology over Z would remove support for the universal-coefficient claim.
Extended reading notes
Core claim
On the paper's own terms, the central claim is Theorem 1: if Σ and Σ' are smooth, oriented, properly embedded surfaces in B^4 with common boundary L ⊂ S^3, and Σ' is obtained from Σ by replacing, in a neighborhood of a curve γ, the standard annulus S^1×B^1 with any smooth annulus A having the same boundary S^1×∂B^1, then for every commutative ring R the maps R ≅ Kh(∅;R) → Kh(L;R) induced by the two surfaces are equal. Rim surgery is a special case, so the theorem answers the open question whether Khovanov cobordism maps can detect rim-surgery exotica: they cannot. The reason is that both maps factor through the skein lasagna module of S^1×B^3 relative to the standard two-component link in it
Load-bearing premise
The load-bearing premise is that Rozansky–Willis homology is functorial over every commutative ring, a result cited from an unpublished thesis; if that functoriality fails, the arbitrary-ring version of the theorem collapses (the first proof covers only fields of characteristic not 2).
Editorial extensions
If this is right
- Khovanov cobordism maps cannot distinguish surfaces in B^4 that differ by a local annulus replacement, so rim surgery is invisible to this invariant over every coefficient ring.
- For every genus g ≥ 0 there exist exotic surface pairs in B^4, with common boundary a knot, that are not related by any sequence of local annulus replacements or rim surgeries.
- The insensitivity extends to any simply connected 4-manifold: two surfaces related by local annulus replacements represent the same element of the skein lasagna module.
- Taking parallel cables does not restore detectability: for any n, the induced Khovanov maps on the n-cables of the two surfaces also agree.
- The equality of maps holds for every commutative ring R, including Z/2, where the first proof's field assumption fails.
Reading between the lines
- Not in the paper: the same cap-off mechanism suggests any link homology theory whose skein module over S^1×B^3 has a rank-one relevant piece will be blind to local annulus replacements; testing this with sl(N) Khovanov–Rozansky homology is a direct next step.
- The theorem indicates that Khovanov-homology detectability of exotic surfaces in the 4-ball is non-local: any invariant that does detect rim-surgery exotica must see the global embedding, not just the surface near a curve.
- A reader who wants to rely on the arbitrary-ring version should first verify the cited functoriality of Rozansky–Willis homology in the unpublished thesis; until then, only the characteristic-not-2 field version stands on the first proof alone.
Signed reviews
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper proves that rim surgery and, more generally, local annulus replacements on a smooth, oriented, properly embedded surface in B^4 do not change the induced map on Khovanov homology from the empty link to the boundary link. Theorem 1 states this for every commutative ring R. Two proofs are offered. The first uses 0-dimensional-input Khovanov skein lasagna modules, works over a field of characteristic not 2, and relies on a computation from Manolescu–Walker–Wedrich. The second uses 1-dimensional-input lasagna modules and Rozansky–Willis homology; this is the proof that extends to arbitrary coefficient rings. The authors also derive a corollary, using work of Hayden–Sundberg, producing exotic pairs of surfaces in B^4 that cannot be related by any sequence of local annulus replacements.
Significance. If the theorem is correct, it answers a question raised by Hayden and Sundberg and demonstrates that Khovanov homology, via skein lasagna modules, has a new insensitivity property for surface surgeries in the 4-ball. The corollary distinguishing exotic pairs by their non-relation under local annulus replacements is a nice application. The paper is clearly organized and the second proof is attractively short, conditional on substantial machinery from the first author's thesis. However, the first proof contains a concrete false topological assertion, and the full-strength arbitrary-ring statement depends on unpublished thesis material. These issues prevent acceptance in the current form.
major comments (3)
- [§2, Step 2] The assertion 'Since γ is contractible in B^4, we may find a smoothly embedded oriented disk D⊂int(B^4) bounding γ' is false. Null-homotopy does not imply the existence of a smooth embedded null-cobordism in dimension 4. For example, let K⊂S^3 be a smoothly non-slice knot (e.g., the trefoil), let Σ be a Seifert surface for K pushed into B^4, and let γ be a parallel copy of K in int(Σ). Then γ is null-homotopic in B^4, but if it bounded a smooth disk D, the annulus in Σ between γ and K would make K slice, a contradiction. Thus the topological reduction to the model (H_{2,n}, vertical disks, spun annulus) is unjustified. Since this reduction is what makes the first proof work, the first proof as written is incomplete.
- [§2, Step 2 (framing conditions)] Even if a disk D bounding γ existed, the argument that D can be chosen so that conditions (i) and (ii) hold is too compressed. The obstructions are said to 'take values in Z' and to be controllable by boundary twists, but no definitions or verifications are given. The sentence 'By the even assumption ... the two obstructions sum to an even integer' does not by itself show that both obstructions can be made to vanish simultaneously; the signs of the effects of a boundary twist on the two obstructions need to be tracked explicitly. This is load-bearing for the same reduction as the previous comment.
- [§3, proof of Theorem 1 over arbitrary R] The proof for arbitrary commutative rings depends entirely on functoriality results stated to be in the unpublished PhD thesis [Ren26, Chapter 4]. The cited preprint [Ren+25] is said to establish only Q-coefficient functoriality; the passage to all rings, and indeed to chain-level homotopy functoriality, is attributed to the thesis, which is not available to the reader. If those results are not public, the universal-coefficient statement in Theorem 1 is not verified by this paper. The authors should either include the necessary statements and proofs, or state Theorem 1 with a coefficient restriction that is actually supported by the public literature.
minor comments (4)
- [§2, Step 1] The phrase 'by linearity, it suffices to prove Theorem 1 when Σ′ is obtained from Σ by a rim surgery (or indeed by a rim surgery with pattern the trefoil knot)' is unclear as it stands. The fact that every annular skein represents a class of the form (1,m(A)) does not, by itself, reduce the general annulus replacement to a single pattern. Since Step 2 later proves the rim-surgery case for every knot K via crossing changes, a more explicit statement of how the reduction is completed would help.
- [§3, notation] The notation KhR^{-,0,0}_{2,0-div}(1_2) is introduced without a definition of the two superscript triples or the '0-div' subscript. The remark about suppressing 'mul' is helpful, but a short explicit definition of the grading convention would make the paper more readable.
- [References] The dependency on [Ren26] is significant; since the thesis is not yet publicly available, the authors should state which specific results are quoted from it (e.g., Propositions/Theorems), or provide a public preprint version of the relevant chapter.
- [§4, Figure 2] The description around the standard diagram of 1_4 would benefit from a more detailed explanation of how the orientations alternate and how the 'dotted and undotted caps' are chosen; Figure 2 is referenced but the surrounding text is terse.
Circularity Check
Load-bearing self-citation for universal-coefficient theorem; no circular derivation otherwise
-
self citation load bearing
[Section 3 (second proof), paragraph on functoriality of Rozansky–Willis homology]
"The first author subsequently removed the restriction on the coefficients and extended the functoriality to any coefficient ring (and indeed to the chain level up to chain homotopy) [Ren26, Chapter 4]."
Theorem 1 is stated for every commutative ring R, but the first proof explicitly assumes a field of characteristic not 2. The universal statement therefore depends entirely on the second proof, whose key input is functoriality of Rozansky–Willis homology over arbitrary coefficient rings. That input is not proved in this paper; it is cited from the first author's own unpublished PhD thesis [Ren26], while the cited preprint [Ren+25] is stated for Q-coefficients. Thus the strongest form of the central claim reduces to a self-citation that is not independently verified or publicly available.
full rationale
The derivation chain of Theorem 1 is not self-referential in the usual circular sense: the maps are factored through Khovanov skein lasagna modules, the local annulus replacement is converted to a rim-surgery case by linearity, and the trefoil rim-surgery case is reduced to the unknot using braid-group symmetry from [GLW18] and crossing changes. No fitted parameter is renamed as a prediction, no equation is equivalent to itself by construction, and no uniqueness theorem is imported from the authors' prior work. The main circularity-adjacent issue is the second proof's reliance, for the 'every commutative ring' claim, on functoriality results cited from the first author's unpublished thesis; this is a load-bearing self-citation rather than an independent check. Separately, the assertion in §2 Step 2 that a contractible curve γ in B^4 bounds a smoothly embedded disk is false in dimension 4 (e.g., a trefoil is null-homotopic but not slice); this is a concrete correctness flaw in the first proof, but it is not a circularity. Weighing the load-bearing self-citation against the independent field-coefficient proof, score 4.
Assumptions & free parameters
assumptions (7)
- domain assumption Functoriality of Khovanov-Rozansky homology for link cobordisms in B^4
- domain assumption Computation of the 0-dimensional-input skein lasagna module S^2_0(S^1×B^3;1_2) ≅ k^4 and the formula for S^1×T_K representing (1,n(K))
- domain assumption Every once-dotted sphere in S^4 evaluates to 1 under KhR_2
- domain assumption Functoriality of Rozansky-Willis homology over arbitrary coefficient rings for relative 1-handlebody complements
- domain assumption The relevant graded piece of the 1-dimensional-input lasagna module S^{2,1d,0-div}_{0,0,0}(S^1×B^3;1_2) is free of rank 1 over R
- domain assumption Grigsby–Licata–Wehrli result: the braid group action on m-cables descends to a symmetric group action on Khovanov homology
- domain assumption The topological obstruction argument for choosing a disk D with prescribed framing conditions
Cite this review
Pith. "Pith review of Insensitivity of Khovanov homology under rim surgery." pith.science (2026). https://pith.science/paper/5IZ2V3BS
@misc{pith2026260725302,
author = {Pith},
title = {Pith review of: Insensitivity of Khovanov homology under rim surgery},
year = {2026},
howpublished = {\url{https://pith.science/paper/5IZ2V3BS}},
note = {Machine review of arXiv:2607.25302}
}
abstract
We show that rim surgery on a smooth, oriented, properly embedded surface in $B^4$ does not change the map on Khovanov homology induced by the surface, answering a question raised by Hayden and Sundber. More generally, the same insensitivity holds for a class of local surgery operations that we call local annulus replacements. Combined with the work of Hayden--Sundberg, this result yields exotic pairs of surfaces in $B^4$ that cannot be related by any sequence of local annulus replacements. We give two proofs, both making essential use of Khovanov skein lasagna modules.
Figures
Reference graph
Works this paper leans on
-
[1]
Khovanov homology and exotic surfaces in the 4-ball , author=. J. Reine Angew. Math. , volume=. 2024 , publisher=
2024
-
[2]
Invariants of 4-manifolds from Khovanov-Rozansky link homology , author=. Geom. Topol , volume=
-
[3]
arXiv preprint arXiv:2510.05273 , year=
Khovanov skein lasagna modules with 1 -dimensional inputs , author=. arXiv preprint arXiv:2510.05273 , year=
-
[4]
arXiv preprint arXiv:2402.10452 , year=
Khovanov homology and exotic 4 -manifolds , author=. arXiv preprint arXiv:2402.10452 , year=
-
[5]
Transverse invariants and exotic surfaces in the 4--ball , author=. Geom. Topol. , volume=. 2021 , publisher=
2021
-
[6]
Surfaces in 4-manifolds , author=. Math. Res. Lett. , volume=. 1997 , publisher=
1997
-
[7]
Skein lasagna modules and handle decompositions , author=. Adv. Math. , volume=. 2023 , publisher=
2023
-
[8]
Annular Khovanov homology and knotted Schur--Weyl representations , author=. Compos. Math. , volume=. 2018 , publisher=
2018
Show all 17 references
-
[9]
arXiv preprint arXiv:1011.1958 , year=
A categorification of the stable SU(2) Witten-Reshetikhin-Turaev invariant of links in S^2 S^1 , author=. arXiv preprint arXiv:1011.1958 , year=
1958 arXiv
-
[10]
Michigan Math
Khovanov homology for links in \#^r(S^2 S^1) , author=. Michigan Math. J. , volume=. 2021 , publisher=
2021
-
[11]
2026 , school=
Khovanov homology and smooth 4-manifolds , author=. 2026 , school=
2026
-
[12]
Smooth surfaces with non-simply-connected complements , author=. Algebr. Geom. Topol. , volume=. 2008 , publisher=
2008
-
[13]
Forum Math
Knotted surfaces in 4-manifolds , author =. Forum Math. , doi =
-
[14]
Modifying surfaces in 4--manifolds by twist spinning , author=. Geom. Topol. , volume=. 2006 , publisher=
2006
-
[15]
A functor-valued invariant of tangles , author=. Algebr. Geom. Topol. , volume=. 2002 , publisher=
2002
-
[16]
Duke Math
Khovanov, Mikhail , title=. Duke Math. J. , number=. 2000 , doi=
2000
-
[17]
2026 , eprint=
Khovanov homology and roll-spun slice disks , author=. 2026 , eprint=
2026
Reviewed August 1, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.