{"id":"395a86ae-5ebb-46a8-a9d1-6ebad37adfa0","arxiv_id":"1908.06453","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Reduced alternating diagrams of links in thickened surfaces and of virtual links minimize crossing number, and any two such diagrams of the same link have the same writhe.","lead":"Two mathematicians prove that the simplest alternating drawings of links on surfaces, and of virtual links, already have the fewest crossings and a fixed twist count. The result settles the first two Tait conjectures for virtual links, using Krushkal's two-variable refinement of the Jones polynomial.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Virtual-link Tait theorems hinge on the imported, unproved assertion that every reduced alternating virtual link diagram has minimal genus; without it, the span comparison in equations (16)-(18) of Section 5 collapses and Theorem 5.3 is unsupported.","rationale":"The central claim is the Kauffman-Murasugi-Thistlethwaite theorem for links in thickened surfaces and virtual links. The surface-link part (Corollary 4.5) is proven within the paper and, after checking the r-parallel argument and Lemma 4.3, appears sound. The virtual-link part (Theorem 5.3) is conditional on the minimal-genus assertion for reduced alternating diagrams, which is cited but not proved. This matches the reader's weakest assumption. The destabilization span inequality in Section 5 is sketched rather than fully proved, but it can be justified: a destabilizing curve is disjoint from the diagram and hence from all states, so no cycle splitting occurs, and each z-factor in the intersection with U is replaced by -A^2-A^{-2}, increasing span by 4 per factor; since U has dimension 2ℓ and state homology is isotropic, at most ℓ factors are replaced per state. Thus the span inequality is not the main risk. The main risk is the external minimal-genus assertion; if it fails, Theorem 5.3 is unsupported. The paper itself flags the missing proof in Section 5. No internal inconsistency in the surface-link proof was found. The verdict CONDITIONAL is appropriate.","tokens_in":24576,"tokens_out":23349,"duration_ms":222678,"concrete_test":"Search Green's virtual knot table through 5 or 6 crossings: for every reduced alternating diagram, compute the genus of its Kamada-Kamada surface and compare with the virtual genus (obtained by Kuperberg's algorithm or an independent destabilization routine). If any reduced alternating diagram has surface genus exceeding the virtual genus, the minimal-genus assertion and Theorem 5.3 are false. A positive result for all listed cases would support but not prove the assertion; to close the dependency, the nullity argument of [BCK19] must be written out for the class of reduced diagrams in Definition 1.5.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 5 (p. 26) asserts without proof that any reduced alternating diagram D of a virtual link L has minimal genus, citing [AARH+19], [AEG+19], and the authors' own unpublished [BCK19]. This assertion is load-bearing: the proof of Theorem 5.3 takes an arbitrary diagram D, destabilizes it to a minimal genus diagram D', and then invokes Kuperberg's Theorem 1.2 to equate span(⟨D''⟩Σ') with span(⟨D'⟩Σ') for a reduced alternating diagram D'' (equations (16)-(18)). If D'' could be represented on a lower-genus surface while preserving the virtual link type, then D'' and D' need not be equivalent in the same thickened surface, the span equality fails, and the inequality n'' ≤ n does not follow. The paper explicitly acknowledges the gap ('There are several ways to prove this') but provides no proof, and the cited references are external preprints. The main surface-link theorem (Corollary 4.5) does not rely on this assertion and appears sound; the concern is limited to the virtual-link extension.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proves a Kauffman-Murasugi-Thistlethwaite theorem for alternating links in thickened surfaces: a connected reduced alternating diagram has minimal crossing number among diagrams on the same surface, and any two reduced alternating diagrams of the same oriented link have the same writhe. The result is stated more generally for adequate diagrams, using Krushkal's two-variable Jones-Krushkal polynomial and its associated homological Kauffman bracket. The main technical ingredients are a state-sum definition of the bracket, an adequacy criterion for diagrams on surfaces, the dual state lemma, and an r-parallel argument adapted from Stong. The paper then applies these results to virtual links, claiming the first and second Tait conjectures for reduced alternating virtual links (Theorem 5.3).","tokens_in":24757,"tokens_out":9893,"duration_ms":102416,"significance":"If the results are correct, this is a substantial extension of classical Tait theory to links in thickened surfaces and to virtual links, and it demonstrates that the Jones-Krushkal polynomial is a useful tool for crossing-number questions. The paper contains a clear state-sum setup, a number of nontrivial worked examples, a table of computations, and an explicit generalization of the dual state lemma. The surface-level main theorem is plausibly sound via the later Corollary 4.5 and the r-parallel argument, and the virtual-link extension is an important intended application. However, the virtual-link theorems currently depend on an unproved minimal-genus assertion imported from external and partially unpublished sources, and the direct proof of Theorem 4.1 has a gap involving the cellularity hypothesis of Lemma 2.10(b).","major_comments":[{"comment":"The proof of Theorem 4.1 applies Lemma 2.10(b) to an arbitrary connected diagram D and concludes that span(⟨D⟩Σ) ≤ 4n + 4 − 4g by way of k(SA) + k(SB) ≤ n + 2 − 2g. Lemma 2.10(b) is stated only for cellularly embedded diagrams, and its proof uses that hypothesis to obtain surjectivity of j∗. Since Theorem 4.1 must apply to arbitrary competitor diagrams, which need not be cellularly embedded, the inequality is not justified as written. I note that Corollary 4.5, obtained through Theorem 4.4 and Lemma 4.3, gives a separate route to the same minimality and writhe statements on a fixed surface and does not use Lemma 2.10(b); the authors should either revise Theorem 4.1 to use that route or explicitly add and justify the cellularity hypothesis where needed.","section":"Section 4, Theorem 4.1"},{"comment":"Theorem 5.3 depends on the unproved assertion that every reduced alternating diagram of a virtual link has minimal genus. The text states 'There are several ways to prove this' and cites [AARH+19], [AEG+19], and the unpublished preprint [BCK19] by two of the current authors, but no proof is supplied. This assertion is load-bearing: equation (17) applies Theorem 2.9 to D′′ on the minimal-genus surface Σ′, and equation (18) uses that equality to derive n′′ ≤ n. If some reduced alternating diagram D′′ could be destabilized to a lower-genus surface while preserving the virtual link type, then span(⟨D′′⟩Σ′) would not be 4(n′′−g′) + 4, and the conclusion n′′ ≤ n would not follow. The virtual-link theorem should either include a proof of the minimal-genus claim or be stated conditionally with the precise external result quoted.","section":"Section 5, page 26, equations (15)-(18)"},{"comment":"Theorem 5.2 for virtual knots invokes Conjecture 5.1 as 'known to be true for virtual knots' and refers to Manturov [Man13], but the manuscript does not state the exact theorem from [Man13] that implies the needed claim. Because this conjecture is what lets the proof reduce a minimal crossing diagram to a minimal genus diagram, the statement should quote the relevant result precisely or prove it, so that Theorem 5.2 is independently verifiable from the cited source.","section":"Section 5, Theorem 5.2"}],"minor_comments":[{"comment":"There is a stray parenthesis in the sentence 'In [AFLT02]), Adams et al. use geometric methods...'; it should read 'In [AFLT02], Adams et al. ...'.","section":"Introduction"},{"comment":"The text says the virtual Hopf link diagram and its states are 'depicted in Figure 8', but Figure 8 shows the virtual trefoil; the intended reference appears to be Figure 13.","section":"Example 3.5"},{"comment":"The statements in Corollary 4.5 are relative to diagrams on one fixed surface Σ; making this explicit in the statements would prevent confusion with stable equivalence or virtual-link diagrams on surfaces of different genera.","section":"Corollary 4.5"},{"comment":"The claim that 'any reduced alternating diagram D of a virtual link L has minimal genus' is asserted for reduced diagrams, while the following sentence says 'any alternating virtual link diagram for L has minimal genus'; these differ if non-reduced alternating diagrams are allowed, so the intended hypothesis should be stated consistently.","section":"Section 5, first paragraph"},{"comment":"The reference [BCK19] is an unpublished preprint by two of the current authors and is cited for a central statement in Section 5; if it remains essential, the relevant theorem from that preprint should be stated in the paper or the preprint should be made publicly available in a citable form.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The main concern for the editor is the status of the virtual-link theorems: Section 5 imports a minimal-genus assertion from an unpublished preprint by the current authors ([BCK19]) and from other preprints, and without a proof the final theorems are conditional. The surface-level result appears repairable by routing through Corollary 4.5 rather than through the flawed application of Lemma 2.10(b) in Theorem 4.1. I would encourage the editor to require that the authors either prove the minimal-genus claim or clearly mark the virtual-link theorem as dependent on it."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper proves the Kauffman-Murasugi-Thistlethwaite theorem for adequate diagrams in thickened surfaces, using Krushkal's homological bracket. That is a real extension: the ordinary Jones polynomial is too weak for alternating virtual knots, so the two-variable invariant is the right tool. The surface-level result (Corollary 4.5) appears sound. The proof via Stong's r-parallel argument avoids the sketchier direct argument in Theorem 4.1. I checked the steps: Lemma 2.6 gives the degree bounds for adequate diagrams, and the writhe invariance follows cleanly. Proposition 2.8 (reduced alternating implies adequate) is a nice observation, and the paper includes useful examples and a table of Jones-Krushkal polynomials.\n\nThe soft spots are real but localized. First, the proof of Theorem 4.1 as written applies Lemma 2.10(b), which requires a cellular embedding, to an arbitrary diagram. The inequality k(SA)+k(SB)≤n+2−2g can fail for non-cellular diagrams. This is not fatal because Corollary 4.5 later gives a valid proof of minimality and writhe invariance for adequate diagrams, so the main surface result stands. But the proof of Theorem 4.1 should be rewritten.\n\nSecond, the virtual-link application is conditional. Section 5 asserts that every reduced alternating diagram of a virtual link has minimal genus, citing two preprints and the authors' own unpublished [BCK19]. This is load-bearing: the span comparison in equations (16)-(18) collapses if the assertion fails. The paper explicitly says 'There are several ways to prove this' and then gives no proof. A referee will need to verify the cited results or ask the authors to include a proof. The destabilization span argument is also sketched; the z-substitution rule and the span-increase bound are plausible but need more detail.\n\nThe virtual Tait conjectures are long-open, so this paper deserves serious attention even though the virtual part is conditional. If the minimal-genus claim turns out to be false, Theorem 5.3 is unsupported, but the surface theorem remains a solid contribution. I would send this to a knowledgeable referee and ask for the virtual part to be tightened. The paper is worth citing and likely worth a reading group session for people in the area.","headline":"The KMT theorem for adequate surface diagrams is a real advance and the surface proof holds up, but the virtual-link Tait conjectures hang on an unproved minimal-genus assertion cited to preprints.","tokens_in":25327,"tokens_out":5510,"would_cite":true,"duration_ms":52557,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["57M25","57M27"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper proves the first and second Tait conjectures for alternating links on surfaces and for virtual links, using a two-variable Jones-type polynomial.","keywords":["Kauffman bracket","Jones-Krushkal polynomial","adequate diagram","alternating link","Tait conjectures","virtual link","thickened surface","crossing number"],"falsifier":"A concrete check: take a reduced alternating diagram of a virtual link and look for a simple closed curve on the surface that is disjoint from the diagram and cuts off a handle; destabilizing along such a curve would produce a diagram of the same virtual link on a lower-genus surface, and if that diagram has fewer crossings, the theorem's minimality claim is false. Computationally, this is a finite search over the small virtual knot tables.","tokens_in":24329,"feed_emoji":"🔗","tokens_out":10223,"duration_ms":95285,"temperature":0.7,"pith_summary":"The paper establishes the analogue of the classical Kauffman–Murasugi–Thistlethwaite theorem for links drawn on surfaces: any connected reduced alternating diagram of a link in a thickened surface has the fewest crossings among all diagrams of that link, and any two reduced alternating diagrams of the same oriented link have the same writhe. The same two conclusions are proved for all adequate diagrams, a wider class that includes every reduced alternating diagram, and are then transferred to virtual links. The engine is a two-variable Jones-type invariant whose span is controlled by the diagram's crossings and the surface's genus, so the crossing-counting argument becomes an inequality between polynomial degrees.","feed_headline":"Tait conjectures hold for alternating surface links","feed_subtitle":"A two-variable Jones polynomial shows reduced alternating diagrams minimize crossings and fix writhe.","key_machinery":"The homological Kauffman bracket $\\langle D\\rangle_\\Sigma$ is a state sum over the $2^n$ smoothings of $D$; a state $S$ contributes $A^{a(S)-b(S)}(-A^{-2}-A^2)^{k(S)}z^{r(S)}$, where $k(S)$ counts the cycles of $S$ that are null-homologous in $\\Sigma$ and $r(S)$ is the rank of the map $H_1(S)\\to H_1(\\Sigma)$. The variable $z$ is what makes the invariant sensitive to the surface, and the Jones-Krushkal polynomial is its normalized, oriented version. Adequacy means that switching one smoothing away from the all-$A$ (or all-$B$) state never increases $k$; Proposition 2.8 shows every reduced alternating surface diagram is adequate, and the dual state lemma bounds $k(S)+k(S^\\vee)$ for cellularly embedded diagrams. These ingredients combine into the span equalities that carry the proof.","core_discovery":"The central claim is that a homological refinement of the Jones polynomial is strong enough to force crossing minimality in surfaces. For a diagram $D$ with $n$ crossings on a surface $\\Sigma$ of genus $g$, the paper proves the span bound $\\operatorname{span}(\\langle D\\rangle_\\Sigma) \\le 4n - 4g + 4$, with equality for adequate diagrams; for a connected reduced alternating diagram, equality holds and the span is exactly $4n - 4g + 4$. Because the span is invariant under Reidemeister moves, any diagram of the same link must have at least $n$ crossings, and a parallel-construction argument shows that two adequate diagrams of the same oriented link have the same writhe. For virtual links, the same results follow once reduced alternating diagrams are known to represent the virtual link on a minimal-genus surface.","pith_inferences":["A natural test of the virtual-link result is the paper's Conjecture 5.1: a computational search for a reduced alternating virtual diagram with a destabilizing curve would either confirm or refute the minimal-genus premise on which the transfer from surfaces to virtual links rests.","The adequacy condition is weaker than the classical plus/minus-adequacy, so the same span machinery should give crossing-number lower bounds for families of non-alternating surface diagrams whose all-$A$ and all-$B$ states can be controlled, such as positive surface diagrams.","Because the Jones-Krushkal polynomial is a homological refinement, a triply graded homology categorifying it would be a natural candidate for an invariant that detects virtual unknots, a task ordinary virtual Khovanov homology is known not to accomplish.","The span equality for reduced alternating surface diagrams has the same shape as the classical Jones-polynomial identity, which suggests asking whether equality on a surface also forces a sequence of flype moves; the paper explicitly leaves the flyping question open for virtual links."],"forward_implications":["For a link in a thickened surface, any connected reduced alternating diagram gives its exact crossing number, so alternating surface links have a computable, diagram-independent crossing number.","Any two reduced alternating diagrams of the same oriented link in a thickened surface have equal writhe, so the writhe is a well-defined invariant for alternating surface links.","The same two conclusions hold for virtual links: reduced alternating virtual diagrams have minimal crossing number, and reduced alternating diagrams of the same virtual link share a writhe.","The minimality result is not limited to alternating diagrams: it applies to every adequate diagram on a fixed surface.","Because reduced alternating diagrams are adequate, the first and second Tait conjectures for surface links are consequences of a single span equality for the Jones-Krushkal polynomial."],"supporting_citations":[{"why":"Defines the homological Kauffman bracket and the Jones-Krushkal polynomial, the invariant whose span carries the argument.","marker":"[Kru11]"},{"why":"Supplies the simple proof of the dual state lemma that the paper adapts to surface diagrams.","marker":"[Tur87]"},{"why":"The r-parallel construction used to prove writhe invariance for adequate diagrams is adapted from this paper.","marker":"[Sto94]"},{"why":"The Kamada-Kamada construction produces the cellular surface diagrams on which the invariants are computed.","marker":"[KK00]"},{"why":"Kuperberg's uniqueness of irreducible representatives lets virtual links be studied on minimal-genus surfaces.","marker":"[Kup03]"},{"why":"Used with the next reference to assert that alternating virtual link diagrams have minimal genus for the link.","marker":"[AARH+19]"},{"why":"Supplies the theorem that tg-hyperbolicity implies minimal genus, one of the two cited routes to that assertion.","marker":"[AEG+19]"},{"why":"Cited as the alternative Gordon-Litherland pairing proof that alternating virtual diagrams are minimal genus.","marker":"[BCK19]"}],"fun_headline_variants":["Surface Tait conjectures proven via two-variable Jones","Alternating surface links minimize crossings and fix writhe","Kauffman-Murasugi-Thistlethwaite extended to thickened surfaces","Crossing minimality for alternating surface diagrams","Adequate diagrams settle Tait conjectures on surfaces"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that every reduced alternating diagram of a virtual link already sits on the lowest-genus surface representing that link, so that destabilization cannot yield a smaller-genus diagram of the same virtual link; if that premise fails, the span inequality that drives the virtual-link proof no longer follows.","fun_headline_variants_meta":{"raw":{"variants":["Surface Tait conjectures proven via two-variable Jones","Alternating surface links minimize crossings and fix writhe","Kauffman-Murasugi-Thistlethwaite extended to thickened surfaces","Crossing minimality for alternating surface diagrams","Adequate diagrams settle Tait conjectures on surfaces"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000479,"raw_usage":{"total_tokens":2300,"prompt_tokens":799,"completion_tokens":1501,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":415,"completion_tokens_details":{"reasoning_tokens":1431}},"tokens_in":415,"tokens_out":1501,"duration_ms":11237,"temperature":1.0,"reasoning_tokens":1431,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T12:46:43.358469+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A concrete check: take a reduced alternating diagram of a virtual link and look for a simple closed curve on the surface that is disjoint from the diagram and cuts off a handle; destabilizing along such a curve would produce a diagram of the same virtual link on a lower-genus surface, and if that diagram has fewer crossings, the theorem's minimality claim is false. Computationally, this is a finite search over the small virtual knot tables.","supporting_citations":[],"review_version":1}