Pith. sign in

REVIEW 3 major objections 8 minor 29 references

$\mathfrak{sl}(2)$-weight system does not extend to a graph 4-invariant

T0 review · 3 major / 8 minor · reviewed 2026-07-31 · grok-4.5

Pith's one-line read The sl(2)-weight system does not extend to a graph 4-invariant in general.

desk verdict Lando’s question gets a clean negative answer via an explicit 9-vertex certificate, with solid constructive follow-ups for the surviving specializations and coefficients. read the letter →

arxiv 2607.24217 v1 pith:NMTLMYLQ submitted 2026-07-27 math.CO math.GT

classification math.COmath.GT MSC 05C3157M2717B10
keywords sl(2)-weightsystem4-invariantintersectiongraphschorddiagramsCasimireigenvaluesChmutov–Varchenkorelationsgraph4T-relations
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

Weight systems on chord diagrams are the algebraic engine behind finite-type knot invariants. A natural question is whether the famous sl(2) weight system, whose values depend only on the intersection graph of a chord diagram, can be rewritten as a function on all graphs that still obeys the same 4T relations. The paper answers no: there is an explicit linear combination of intersection graphs that evaluates to a nonzero polynomial under the weight system yet becomes zero once ordinary graph 4T relations are imposed. The same calculation isolates four special values of the Casimir parameter at which an extension might still exist; two of those extensions were already known, a third is given by a new recurrence, and the fourth is verified computationally up to ten vertices. The same methods settle which coefficients of the weight-system polynomial extend and which do not. The result shows that factoring through intersection graphs is not enough to guarantee a 4-invariant on the whole graph category.

What carries the argument

The computer-generated certificate C in the span of intersection graphs: a sparse dependence among 4T-relations on nine-vertex graphs whose evaluation by the Chmutov–Varchenko recurrence is the nonzero polynomial above. That single algebraic identity simultaneously disproves general extendability and pins down the only admissible specializations.

What would settle it

Re-run the nine-vertex 4T linear system (or inspect the published certificate) and check whether the combination of intersection graphs really evaluates to c(c−3/8)(c−1)(c+3/32) under the chord-deletion formula and is identically zero in the graph 4T quotient.

Watch

Extended reading notes

Core claim

The sl(2)-weight system does not extend to any 4-invariant of graphs. An explicit certificate—a linear combination of 3300 graph 4T-relations that expands to 5006 intersection graphs—evaluates under the weight system to the nonzero polynomial c(c−3/8)(c−1)(c+3/32). Consequently only the four roots of that polynomial remain candidates for specializations that could extend, and the paper constructs or verifies extensions at three of them while proving that all but the leading few polynomial coefficients fail to extend.

Load-bearing premise

The computer certificate for nine-vertex graphs is complete and free of isomorphism or arithmetic error; if the sparse linear algebra missed a relation or mis-evaluated a graph, the non-extendability claim would collapse.

Editorial extensions

If this is right

  • Only the four Casimir eigenvalues 0, 3/8, 1 and −3/32 can possibly admit graph 4-invariant extensions of the sl(2) weight system.
  • The specializations at c=0, 3/8 and 1 now possess explicit or recurrent graph formulae; the oscillator value c=−3/32 is unique at least through ten vertices.
  • All polynomial coefficients [c^{n−k}] for k ge5 and all constant-term coefficients [c^k] for k ge1 fail to extend to 4-invariants.
  • Even when a weight system factors completely through intersection graphs it need not lift to a 4-invariant on the larger graph category.

Reading between the lines

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

  • The same certificate technique can be applied verbatim to other Lie-algebra weight systems to decide extendability without first constructing candidate formulae.
  • A closed combinatorial formula for the oscillator specialization, if it exists, would complete the dictionary between Casimir eigenvalues and graph invariants.
  • Uniqueness of the three known extensions remains open and may require a separate generating-function or Hopf-algebra argument beyond the range of finite computation.
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 / 8 minor

Summary. The paper answers Lando's long-standing question in the negative: the sl(2)-weight system, although its values depend only on intersection graphs (Chmutov–Lando), does not extend to a 4-invariant of graphs. The obstruction is an explicit n=9 certificate: a linear combination of 3300 graph 4T-relations expanding to 5006 intersection graphs whose Chmutov–Varchenko evaluation is the nonzero polynomial c(c−3/8)(c−1)(c+3/32). Its roots restrict possibly extendable specializations to c ∈ {0, 3/8, 1, −3/32}. On the positive side, the authors construct an explicit extension ψ at c=1 (the 3-dimensional representation) via the Z/2-corank η, prove its 4-invariance and a deletion-type recurrence (Theorem 3.6); conjecture an extension at c=−3/32 related to the oscillator representation, supported by a unique extension computed for graphs on ≤10 vertices; and completely settle the extension question for the polynomial coefficients of w_sl(2): Theorem 5.1 and Corollary 5.2 show [c^{n−k}] extends exactly for k≤4, while [c^k] for k≥1 does not, and Section 5.1 derives an exact 39-term closed formula for [c^{n−3}] by inverting a unitriangular subgraph-counting feature matrix. The deductive skeleton is clean throughout; the central result is computational, resting on the posted certificate.

Significance. This resolves a well-known open problem in finite-type knot invariant theory, posed by Lando and reviewed in [17]: even though the sl(2)-weight system factors through intersection graphs (Chmutov–Lando), it does not lift to a graph 4-invariant. The result is sharp — the certificate polynomial's roots are exactly the four Casimir eigenvalues at which extensions can exist, and the paper treats all four: trivial at c=0, known at c=3/8, a new proved extension at c=1, and a well-supported conjecture at c=−3/32. Strengths worth naming: the obstruction certificate is concrete, publicly posted, cheap to re-verify (a single-point evaluation settles the headline claim), and reportedly independently verified by M. Kazarian; the authors' code reproduces Krasilnikov's n≤8 table; the c=1 recurrence (Theorem 3.6) is proved by hand; the coefficient dichotomy (Corollary 5.2) and the exact 39-term formula for [c^{n−3}] are parameter-free and falsifiable; and the c=−3/32 conjecture comes with falsifiable n≤10 data and a conjectural recurrence. No free parameters or ad hoc axioms are introduced.

major comments (3)
  1. [§2, Proposition 2.1] §2, proof of Proposition 2.1: the entire proof of the paper's central result is the sentence "The certificate can be found at [GitHub URL]." Since the negative answer to Lando's question rests on this object alone, the manuscript should (i) describe the certificate format and size in the text; (ii) state explicitly that the repository contains a self-contained verifier performing the three cheap checks — each of the 3300 rows is a valid graph 4T-relation on correctly identified isomorphism classes, the signed sum equals the claimed combination of 5006 intersection graphs, and its evaluation under recursion (1) is exactly c(c−3/8)(c−1)(c+3/32); and (iii) confirm that every hash match in the deduplication step is confirmed by an explicit isomorphism test, since the degree/2-degree hash of §2, step (1) is not a complete invariant and an unconfirmed collision merging non-isomorphic graphs co
  2. [§3.2, Propositions 3.4–3.5] §3.2, Propositions 3.4–3.5: the 4-invariance argument for ψ is compressed to four sentences. The claim that the inner sum "splits into four framed 4T relations on 3^{η(G')}" requires matching the sign (−1)^b in Definition 1.7 against the (−1)^{|V'|} weights and the framing flips on the two moving vertices; this sign bookkeeping is exactly where such arguments fail, and it is currently left to the reader. Likewise Proposition 3.5 asserts in one line the identification of ψ with the c=1 specialization via Theorem 3.1 and Proposition 3.2 (σ ↔ framing, |σ|−1 ↔ η, ∏σ(c) ↔ (−1)^{|V'|}). Please expand both steps; the c=1 extension is one of the paper's three positive results and deserves a fully written proof comparable to Theorem 3.6.
  3. [§5, Corollary 5.2] §5, Corollary 5.2 (the k>8 and [c^k], k≥1 cases): the proof applies the "add an isolated vertex" and "add a leaf" transforms "to the original certificate," but only the effect on w_sl(2)-values (multiplication by c and by c−1/2) is argued. What is missing is one explicit justification that these operations, applied termwise to a linear combination of graph 4T-relations, again produce a combination of graph 4T-relations (locality of the 4T relation) whose expansion remains a combination of intersection graphs. A short lemma stating this formally would close the non-extension half of the coefficient dichotomy announced in the abstract; as written, that half relies on an unstated compatibility step.
minor comments (8)
  1. [§2] §2: the subscript notation is inconsistent — the algorithm is described "for each n_i" and terminates "after all n_i = 0,…,n have been processed," but n_i is never defined; presumably these are just the integers 0,…,n.
  2. [§2, step (1)] §2, step (1): in the definition of the hash, "path of degree exactly 2" should presumably read "path of length exactly 2."
  3. [§5.1, Theorem 5.7] Theorem 5.7: the proof of the 39-term closed formula for [c^{n−3}] is again only a URL. Given that Lemmas 5.4–5.5 reduce it to an exact inversion of a 208-dimensional unitriangular matrix, please include in the text (or an appendix) the dimension and a spot check — e.g., evaluate both sides on several graphs on 7–8 vertices using the convolution of Theorem 5.1 — so the formula does not rest solely on the pipeline artifact.
  4. [§5.1, footnote] The footnote in §5.1 reconciling the quadrangle conventions (qcd = −2×…) is hard to parse; a short displayed example matching one chord-diagram quadrangle to its three induced graph types would help.
  5. [§3.3, Conjecture 3.1] Conjecture 3.1: please state whether the verification on graphs with ≤9 vertices used exact rational arithmetic or floating point; the same question applies to the n≤10 table for Conjectures 4.1–4.2.
  6. [§1.2] §1.2, Remark: since 1T-relations are not imposed, state explicitly that Question 1.1 is answered in the framed sense (consistent with [9, 19]); a sentence on whether imposing 1T would affect the obstruction would be welcome.
  7. [Table 1] Table 1: add a source or method note for the values dim I_9 = 127954 and dim I_10 = 2165291, which to my knowledge are not tabulated elsewhere.
  8. [§1.4, Definition 1.8] Definition 1.8: clarify that the symbol "a" in "preserved when a=0 and switched when a=1" refers to the framing of vertex a, not the vertex itself.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the negative answer is an external linear-algebra obstruction, not a result forced by its own inputs.

full rationale

The load-bearing claim (Proposition 2.1) is that there exists C in the span of intersection graphs with w_sl(2)(C) ≠ 0 whose image in G/4T is zero. That C is produced as a concrete linear combination of graph 4T-relations; its non-vanishing is the evaluation of the independently axiomatized Chmutov–Varchenko weight system (Definition 1.4, chord-deletion formula (1)) on that combination, yielding the nonzero polynomial c(c−3/8)(c−1)(c+3/32). Nothing in that chain defines w_sl(2) from 4T-vanishing or fits a parameter to force the obstruction. Specialization values are identified with known Casimir eigenvalues of sl(2) representations, not tuned to the certificate. The [c^{n−3}] closed form (Theorem 5.7) is an exact change of basis in a lower-unitriangular subgraph-counting feature matrix (Lemma 5.4), explicitly not a statistical prediction. Self-citations ([12], GitHub artifacts, Krasilnikov reproduction) supply prior extensions or verification data and are not used as uniqueness theorems that forbid alternatives. Conjectures (oscillator representation, Conjectures 3.1 and 4.1–4.2) are labeled as such. The derivation is therefore self-contained against its stated inputs; residual risk is ordinary computational correctness of the public certificate, not internal circularity.

Assumptions & free parameters 0 free parameters · 5 assumptions · 3 invented entities

The paper works inside the standard Chmutov–Varchenko axiomatization of the sl(2)-weight system, the intersection-graph map, and Lando’s 4T-relations on graphs. No free numerical parameters are fitted to obtain the negative answer. Invented objects are the explicit certificate C and the graph function ψ; both are defined constructively. Conjectural oscillator extension ξ is clearly separated from proved claims.

assumptions (5)
  • domain assumption Chmutov–Varchenko axioms uniquely define w_sl(2):A→C[c] (normalization, multiplicativity, leaf deletion, 6T), implying 4T.
    Taken as the definition of the weight system throughout Sections 1.3 and 2; standard in the literature [8].
  • domain assumption Graph 4T-relations (and 2T transforms) define 4-invariants on (framed) graphs; intersection graphs send diagram 4T to graph 4T.
    Definition 1.7–1.8 and the commuting diagram in Figure 1; background from Lando et al.
  • standard math η(G)=corank of the Z/2 adjacency matrix is a 2T-invariant extending |σ|−1 on diagrams.
    Cited from [23,10] and used to build ψ at c=1 (Proposition 3.2).
  • domain assumption Casimir eigenvalues 0, 3/8, 1, −3/32 correspond to the 1-dim, 2-dim irreducible, 3-dim irreducible, and oscillator representations of sl(2).
    Used in Section 2 to interpret the roots of the certificate polynomial; standard representation theory plus the oscillator construction in Section 4.
  • standard math Convolution of 4-invariants is a 4-invariant; [c^{n−k}] on diagrams is 1_{n−2k}·f_{2k} (Chmutov–Varchenko Thm 3).
    Invoked for Theorem 5.1 extending coefficients from ≤2k vertices to all graphs.
invented entities (3)
  • Certificate C (linear combination of ~3300 graph 4T-relations / 5006 intersection graphs on 9 vertices) independent evidence
    purpose: Explicit obstruction showing w_sl(2) does not descend to G/4T.
    Constructed computationally in Section 2; value c(c−3/8)(c−1)(c+3/32) forces the negative answer and the candidate list.
  • Graph function ψ(G)=2^{−|V|} ∑_{V'} (−1)^{|V'|} 3^{η(G')} independent evidence
    purpose: 4-invariant extending w_sl(2) at c=1 (3-dimensional representation).
    Definition 3.3; proved 4-invariant and matching on diagrams; recurrence in Theorem 3.6.
  • Conjectural 4-invariant ξ at c=−3/32 (oscillator specialization)
    purpose: Candidate unique extension for the last remaining Casimir eigenvalue.
    Conjecture 4.1–4.2; uniqueness verified only through 10 vertices; no closed formula.

how reviews work

0 comments
Cite this review

Pith. "Pith review of $\mathfrak{sl}(2)$-weight system does not extend to a graph 4-invariant." pith.science (2026). https://pith.science/paper/NMTLMYLQ

@misc{pith2026260724217,
  author       = {Pith},
  title        = {Pith review of: $\mathfraksl(2)$-weight system does not extend to a graph 4-invariant},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/NMTLMYLQ}},
  note         = {Machine review of arXiv:2607.24217}
}
abstract

A long-standing question by S. Lando asks whether the $\mathfrak{sl}(2)$-weight system extends to a unique 4-invariant of graphs. We show that, in full generality, the answer to this question is negative. However, for certain specializations of the weight system, extensions do exist. Explicit formulae for computing two such specializations of the weight system are already known. We construct recurrence relations for one additional such extension and discuss the last remaining specialization, which conjecturally admits an extension. We also study the polynomial coefficients of the $\mathfrak{sl}(2)$-weight system and resolve the questions concerning their extension.

Figures

Figures reproduced from arXiv: 2607.24217 by the authors.

Figure 1
Figure 1. Lando’s question: are there unique g, gb making the diagram commute? We answer the question of Lando in the negative by providing such an element (a certificate) C ∈ I that wsl(2)(C) ̸= 0 and the image of C in G/ 4T is 0. The calculated value wesl(2)(C) restricts the set of values of the variable c for which the corresponding evaluation of the sl(2)-weight system can possibly admit extension to G/ 4T to the roots of… view at source ↗
Figure 2
Figure 2. The product of chord diagrams is unique modulo 4T-relations [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. Example of the intersection graph of a chord diagram. †Any other choice of the metric on sl(2) leads to a controllable rescaling of the coefficients of the sl(2)-weight systems. See [8, 4] for the details [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: If we only care about the neighborhood of vertex v of a graph provided with a vertex grouping, we abbreviate it as illustrated. Here vertex w is outside of the region of interest so we choose to omit it in the abbreviated figure. We will make use of adjacency matrices …
Figure 5
Figure 5. Figure 5: FG≤3 . Write fj (Gi) := Fi,j for the j-th feature of the graph Gi , i.e. the number of subgraphs of Gi isomorphic to Gj . For each n, let G≤n be the vector space spanned by G≤n with basis e = {Gj | Gj ∈ G≤n} and G∗ ≤n its dual space of linear functions. Each feature fj…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

29 extracted references · 2 linked inside Pith

  1. [17]

    Maxim Kazaryan and Sergei Lando.Weight systems and invariants of graphs and embedded graphs. 2023. arXiv:2302.12153 [math.CO]

  2. [1]

    On the Vassiliev knot invariants

    Dror Bar-Natan. “On the Vassiliev knot invariants”. In:Topology34 (1995), pp. 423–472

  3. [2]

    Dror Bar-Natan.Weights of Feynman diagrams and the Vassiliev knot invariants. 1991

  4. [3]

    Proof of a conjecture of Kulakova et al. related to thesl2 weight system

    Dror Bar-Natan and Huan T. Vo. “Proof of a conjecture of Kulakova et al. related to thesl2 weight system”. In:European Journal of Combinatorics45 (2015), pp. 65–70

  5. [4]

    Chmutov, S

    S. Chmutov, S. Duzhin and J. Mostovoy.Introduction to Vassiliev knot invariants. 2012

  6. [5]

    Vassiliev knot invariants I. Introduction

    S. V. Chmutov, S. V. Duzhin and S. K. Lando. “Vassiliev knot invariants I. Introduction”. In:Advances in Soviet Mathematics21 (1994), pp. 117–126

  7. [6]

    Vassiliev knot invariants II. Intersection graph conjecture for trees

    S. V. Chmutov, S. V. Duzhin and S. K. Lando. “Vassiliev knot invariants II. Intersection graph conjecture for trees”. In:Advances in Soviet Mathematics21 (1994), pp. 127–134

  8. [7]

    Vassiliev knot invariants III. Forest algebra and weighted graphs

    S. V. Chmutov, S. V. Duzhin and S. K. Lando. “Vassiliev knot invariants III. Forest algebra and weighted graphs”. In:Advances in Soviet Mathematics21 (1994), pp. 135–145

Show all 29 references
  1. [8]

    Remarks on the Vassiliev knot invariants coming fromsl2

    S. V. Chmutov and A. N. Varchenko. “Remarks on the Vassiliev knot invariants coming fromsl2”. In:Topology 36.1 (1997), pp. 153–178

  2. [9]

    Mutant knots and intersection graphs

    Sergei Chmutov and Sergei Lando. “Mutant knots and intersection graphs”. In:Algebraic & Geometric Topol- ogy7.3 (2007), pp. 1579–1598

  3. [10]

    Cycle decomposition by disjoint transpositions

    M. Cohn and A. Lempel. “Cycle decomposition by disjoint transpositions”. In:Journal of Combinatorial Theory, Series A13.1 (1972), pp. 83–89

  4. [11]

    On chromatic coefficients

    E. J. Farrell. “On chromatic coefficients”. In:Discrete Mathematics29 (3 1980), pp. 257–264

  5. [12]

    Thesl(2)-weight system atc= 3/8for graphs

    Daniil Fomichev and Maksim Karev. “Thesl(2)-weight system atc= 3/8for graphs”. In:European Journal of Combinatorics127 (2025)

  6. [13]

    2006.url:https : / / web

    Paul Garrett.The Segal–Shale–Weil (oscillator) representation. 2006.url:https : / / web . archive . org / web/20210827124200/https://www-users.cse.umn.edu/~garrett/m/v/oscillator_repn.pdf(visited on 27/08/2021)

  7. [14]

    Humphreys.Representations of Semisimple Lie Algebras in the BGG CategoryO

    James E. Humphreys.Representations of Semisimple Lie Algebras in the BGG CategoryO. Graduate Studies in Mathematics 94. American Mathematical Society, 2008

  8. [15]

    On the primitive subspace of Lando framed graph bialgebra

    Maksim Karev. “On the primitive subspace of Lando framed graph bialgebra”. In:Communications in Math- ematics31.3 (2024)

  9. [16]

    The space of framed chord diagrams as a Hopf module

    Maksim Karev. “The space of framed chord diagrams as a Hopf module”. In:Journal of Knot Theory and Its Ramifications24.03 (2015), p. 1550014

  10. [18]

    Vassiliev’s knot invariants

    Maxim Kontsevich. “Vassiliev’s knot invariants”. In:Advances in Soviet Mathematics16.2 (1993), pp. 137– 150

  11. [19]

    An Extension of thesl 2 Weight System to Graphs withn≤8Vertices

    Evgeny Krasilnikov. “An Extension of thesl 2 Weight System to Graphs withn≤8Vertices”. In:Arnold Mathematical Journal7 (2021), pp. 609–618

  12. [20]

    On a weight system conjecturally related tosl2

    E. Kulakova et al. “On a weight system conjecturally related tosl2”. In:European Journal of Combinatorics 41 (2014), pp. 266–277

  13. [21]

    J-invariants of plane curves and framed chord diagrams

    Sergei Lando. “J-invariants of plane curves and framed chord diagrams”. In:Functional Analysis and Its Applications40.1 (2006), pp. 1–10

  14. [22]

    On a Hopf algebra in graph theory

    Sergei K. Lando. “On a Hopf algebra in graph theory”. In:Journal of Combinatorial Theory, Series B80.1 (2000), pp. 104–121. 14 REFERENCES

  15. [23]

    Lando and Alexander K

    Sergei K. Lando and Alexander K. Zvonkin.Graphs on Surfaces and Their Applications. Encyclopaedia of Mathematical Sciences. Springer Berlin, Heidelberg, 2010

  16. [24]

    Read and Robin J

    Ronald C. Read and Robin J. Wilson.An Atlas of Graphs. Oxford University Press, 1998

  17. [25]

    Cohomology of knot spaces

    V. Vassiliev. “Cohomology of knot spaces”. In:Advances in Soviet Mathematics1 (1990), pp. 23–70

  18. [26]

    Proof of a conjecture of Fomichev and Karev

    Qi Yan, Qingying Deng and Xian’an Jin. “Proof of a conjecture of Fomichev and Karev”. In:European Journal of Combinatorics135 (2026)

  19. [27]

    Polina Zakorko and Polina Zinova.Duality for thesl2 weight system. 2024. arXiv:2407.01144 [math.CO]

  20. [28]

    Значенияsl 2-весовой системы на семействе графов, не являющихся графами пере- сечений хордовых диаграмм

    П. А. Филиппова. “Значенияsl 2-весовой системы на семействе графов, не являющихся графами пере- сечений хордовых диаграмм”. In:Матем. сб.213.2 (2022), pp. 115–148

  21. [29]

    Значения весовой системы, отвечающей алгебре Лиsl 2, на полных двудольных графах

    П. А. Филиппова. “Значения весовой системы, отвечающей алгебре Лиsl 2, на полных двудольных графах”. In:Функц. анализ и его прил.54.3 (2020), pp. 73–93. D.F.: Saint-Petersburg State University, 199034, 7/9 Universitetskaya Emb., Saint-Petersburg, Russia Email address:fomichev....

Pith tools

Reviewed July 31, 2026 · model on record in the stance chip above.