{"id":"3882c56a-3aad-40dd-84c9-0fac0d918e7d","arxiv_id":"2506.15204","paper_version":1,"verdict":"REJECT","confidence":"HIGH","novelty_score":4.0,"correctness_risk":"high","formal_verification":"none","parameter_count":0,"one_line_summary":"The paper claims new split exact sequences for SK of pairs with tangential structures and maps, plus a product theorem SK_n(X×Y) ≅ SK_n(X) for simply connected Y, but a central claimed example of a strongly multiplicative structure is false.","lead":"This paper studies cutting and pasting groups of pairs of manifolds with tangential structures, claiming to extend Komiya's pair splitting results to manifolds with maps into a reference space. It also proves a product splitting theorem for oriented SK-groups when one factor is simply connected.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Example 2.6(3) is valid; the real gap is the 4D proof of Theorem 6.1, on which Theorem 3.3 depends.","rationale":"The reader's REJECT rests on a mistaken reading of Example 2.6(3): the space B×X is strongly multiplicative by carrying the X-coordinate of the second factor through the map ψ_{r,s}. This removes the alleged fatal flaw in Theorems 3.2 and 3.3. However, a careful reading of Section 6 shows that Theorem 6.1 for 4-manifolds is only sketched: the crucial steps involving extensions of maps to Y over surgery traces and the preservation of maps to X under the stable diffeomorphism are not justified. Because Theorem 3.3 uses the projection isomorphism SK_n(X×BSO(m−n)×X)→SK_n(X×X) for all n, including n = 4, the incompleteness of the 4D proof of Theorem 6.1 is a load-bearing gap. The verdict should move from REJECT to CONDITIONAL: the main framework appears sound, but the oriented splitting theorem and Theorem 6.1 require a completed proof in dimension 4 before acceptance.","tokens_in":10226,"tokens_out":35983,"duration_ms":339165,"concrete_test":"Require a complete proof of Theorem 6.1 for n = 4: (i) show the surgeries that make f'_* an isomorphism admit extensions of the maps to Y over the traces; (ii) prove that the stable diffeomorphism can be chosen to intertwine the maps to X×Y up to the SK_4 equivalence, not just the normal 1-smoothing; (iii) justify that [N',g''×k] = [N,g'''×k'] in SK_4(X×Y) follows from a bordism whose map to Y is constructed. If these steps cannot be supplied, Theorem 3.3 for n = 4 remains unsupported.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The reader's objection to Example 2.6(3) does not land. For C = B×X, define Ψ((b_r,x_r),(b_s,x_s)) = (ψ_B(b_r,b_s), x_s). The strong-multiplicativity square is a pullback because the target's X-coordinate in the second factor recovers x_s, the first factor's X-coordinate is x_r, and the B-coordinates are determined by the B pullback. So the splitting theorems 3.2 and 3.3 do not collapse. The genuine soft spot is the proof of Theorem 6.1 for n = 4 in Section 6. It is asserted without proof that the surgery traces used to make f'_* and g'_* isomorphisms can be equipped with maps to Y, and that the stable diffeomorphism φ of Theorem 6.2 between M'#aS and N'#bS preserves the maps to X×Y up to the equivalence needed in SK_4. Extending maps over 5-dimensional cobordisms can be obstructed by π_2(Y), and the normal 1-smoothing only records π_1(X), not the full map to X. Since Theorem 3.3's projection isomorphism SK_n(X×BSO(m−n)×X)→SK_n(X×X) invokes Theorem 6.1 for Y = BSO(m−n) in all dimensions, including n = 4, this gap is load-bearing.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies cutting and pasting (SK) groups for pairs (M,N) of manifolds equipped with tangential B-structures. After introducing the notion of a strongly multiplicative structure, it proves Theorem 3.1, a split short exact sequence relating SK^B_m, SK^B_{m,n}, and SK^B_n(B_{m-n}). This is then applied to obtain Theorem 3.2 for unoriented manifolds with maps to a space X and Theorem 3.3 for oriented manifolds with maps to X. The oriented result relies on Theorem 6.1, which asserts that if Y is simply connected and π_1(X) is finitely presented, then the projection SK_n(X×Y)→SK_n(X) is an isomorphism for all n. The proof of Theorem 6.1 is attempted using the asymmetric signature for n≥6 and Kreck's modified surgery for n=4.","tokens_in":10401,"tokens_out":36823,"duration_ms":353072,"significance":"The splitting theorem for pairs is a natural extension of Komiya's work, and the strong-multiplicativity formalism is appropriate. I also note that the claim in Example 2.6(3) that B×X is strongly multiplicative is correct; the required pullback is obtained by setting Ψ((b_r,x_r),(b_s,x_s))=(ψ_B(b_r,b_s),x_s), and the X-coordinates do not obstruct the pullback property. If Theorem 6.1 were fully justified, Theorems 3.3 and B would be useful generalizations of Neumann's theorem. However, the proof of Theorem 6.1 has significant gaps in the current version, particularly in the n=4 case, so the paper's central implications are not yet established.","major_comments":[{"comment":"The step beginning 'Because of this, we have that f' gives a choice of normal 1-smoothing...' is incomplete. Kreck's Theorem 6.2 produces a stable diffeomorphism φ between M'#aS and N'#bS from bordant normal 1-smoothings, but those smoothings only record maps to Bπ_1(X)×BSO, not the full maps to X×Y. The subsequent equality [M,f×h]=[M',f'×h']=[N',g''×k]+Σ[S,q_i] and the claim that 'reversing the surgeries from N to N' is not obstructed by the map g'' because the only obstructions live in the fundamental group' require that the restrictions of the maps extend over the 5-dimensional surgery traces from N' back to N. Extending maps over such cobordisms can be obstructed by π_2(X) and π_2(Y), so this assertion is not justified. This gap is load-bearing because Theorem 6.1 is used in Theorem 3.3, and no alternative argument is supplied.","section":"Section 6, proof of Theorem 6.1 for n=4"},{"comment":"The proof asserts 'σ_*(M,f×g) equals σ_*(M,f)=0 because [M,f] is bordant to a manifold that has an open book decomposition.' The asymmetric signature σ_* is a bounded book bordism invariant, not an ordinary bordism invariant, and an ordinary bordism from M to an OBD manifold does not by itself give a zero class in BB_n(X). The argument needs to justify that vanishing in SK_n(X) implies vanishing of σ_*(M,f), for example by proving or citing that the class of [M,f] is zero in BB_n(X) or by constructing a bounded book null-bordism from the cut-and-paste relation. Without this, the high-dimensional proof of Theorem 6.1 is incomplete.","section":"Section 6, high-dimensional case (n≥6)"}],"minor_comments":[{"comment":"In the proof of Theorem 3.3, 'X×BO(m−n)×X' should read 'X×BSO(m−n)×X'; as printed it is inconsistent with the use of Theorem 6.1 in Section 6.","section":"Theorem 3.3, proof"},{"comment":"In the n=4 proof of Theorem 6.1, 'similarly for h' on N'' should presumably be 'similarly for g' on N''; the map h was only introduced for M.","section":"Section 6, n=4 proof"},{"comment":"The label 'id×φ_{r,s}' for the bottom horizontal map is confusing; the intended map is (a,b)↦(a,φ_{r,s}(a,b)). Please spell this out.","section":"Definition 2.5"},{"comment":"The notation j([(M,N),(f,g)]) is introduced without specifying the domains of f and g; state that f:M→X and g:N→X.","section":"Theorems 3.2 and 3.3"},{"comment":"The sentence 'In odd dimensions all the SK-groups vanish' is used without proof or reference; please supply a reference or a short justification.","section":"Section 6, low-dimensional cases"}],"recommendation":"major_revision","confidential_remarks":"The paper's Theorem 3.1 and the unoriented application appear sound, but the proof of Theorem 6.1, especially the n=4 case, is substantially under-supported. Since Theorem 3.3 and Theorem B depend on Theorem 6.1, the author should either complete the missing arguments or restrict the claims to the cases that are proved. I would advise sending the revised version to an expert in Kreck's modified surgery to assess the n=4 argument."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Things you should know before spending time on arXiv:2506.15204.\n\nThe reader's rejection reasons are mostly off-target. Example 2.6(3) is actually fine: for the structure B×X, set Ψ((b_r,x_r),(b_s,x_s)) = (ψ_B(b_r,b_s), x_s). The square in Definition 2.5 is a pullback because the X-coordinate in the target comes from the second factor, and the B-coordinate is uniquely determined by strong multiplicativity of B. No map X×X→X is forced. So Theorems 3.2 and 3.3 do not collapse on that ground.\n\nWhat the paper does well: Theorem 3.1 is a clean generalization of Komiya's pair-splitting theorem to strongly multiplicative tangential structures. The Section 2 framework is careful, and the proof of Theorem 3.1 is essentially Komiya's, which is fine. The unoriented Theorem 3.2 follows from the Künneth isomorphism and the known computation of SK^O_*(BO(k)); it looks correct. Theorem 6.1 is a genuine extension of Neumann's theorem, and the high-dimensional proof via Ranicki's asymmetric signature is plausible, modulo the usual faith in that machinery.\n\nThe real soft spot is the n=4 case of Theorem 6.1. The proof there is too compressed. The critical step is where a stable diffeomorphism φ: M'#aS → N'#bS, obtained from Kreck's modified surgery, is used to identify [M'#aS, (f'×h')#a*] with [N'#bS, (f'×h')#a*∘φ^{-1}] in SK_4(X×Y). Kreck's theorem gives φ up to homotopy commuting with the normal 1-smoothings, but those smoothings only remember the map to K(π1(X),1) (via X) and the map to BSO. They do not encode the actual map to X or Y. Nothing in the argument shows that φ preserves the map to X×Y in the sense needed for an SK_4 equality. The sentence 'the only obstructions live in the fundamental group' is doing a lot of work and is not justified; π_2(Y) can obstruct extending maps over 5-dimensional cobordisms. Since Theorem 3.3 invokes Theorem 6.1 with Y=BSO(m−n) for all n, this gap affects a headline result.\n\nBottom line: the paper is worth a serious referee. Theorem 3.1 is a solid contribution for the SK-group community. Theorem 6.1 is attractive but the 4D proof needs to be either completed or the theorem stated with n≠4. I would send it to peer review, with a referee who knows Kreck surgery and Ranicki's asymmetric signature.","headline":"The reader's main objection to Example 2.6(3) is incorrect; the real problem is the underproved n=4 case of Theorem 6.1, which also undercuts Theorem 3.3 as stated.","tokens_in":11040,"tokens_out":9984,"would_cite":true,"duration_ms":79275,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["57R90","57R15"],"pacs":[],"model":"deepseek-v4-flash","headline":"For a large class of tangential structures, cutting and pasting a pair of manifolds is exactly the separate cutting and pasting of the ambient manifold and of the submanifold carrying its normal-bundle structure.","keywords":["SK-groups","cutting and pasting","pairs of manifolds","tangential structures","strongly multiplicative","bordism groups","open book decompositions","stable diffeomorphism"],"falsifier":"Check Example 2.6(3) for $X=S^1$, $r=s=1$. The pullback condition in the definition of strong multiplicativity requires an explicit comparison map out of $(BO_1\\times S^1)\\times(BO_1\\times S^1)$ to be a homeomorphism onto the fibre product determined by the bottom maps; writing both spaces as homotopy equivalent CW complexes and comparing $\\pi_1$ (or $\\pi_2$) is a finite check. If the square is not a pullback, the inheritance claim used for the normal-bundle structure of the submanifold is false; if it is, the claim survives this test.","tokens_in":9879,"feed_emoji":"✂️","tokens_out":20225,"duration_ms":189750,"temperature":0.7,"pith_summary":"Cutting-and-pasting groups (SK-groups) record when two manifolds are equivalent up to cutting along a codimension-one submanifold and regluing the pieces by diffeomorphisms. This paper asks what happens when the object being cut and pasted is a pair: a manifold together with a lower-dimensional submanifold. It proves that, for any tangential structure satisfying a two-out-of-three rule (called strongly multiplicative), the SK-group of pairs splits into the SK-group of the ambient manifold and an SK-group of the submanifold equipped with its normal-bundle structure. As corollaries, manifolds with a map into a reference space $X$ yield analogous splittings, and maps into a simply connected factor $Y$ of a product $X\\times Y$ are invisible to oriented SK-groups. The upshot is that many pair-SK computations reduce to known single-manifold computations, making the pair theory substantially more tractable.","feed_headline":"Pairs of manifolds split into two cutting-and-pasting groups","feed_subtitle":"For strongly multiplicative structures, the submanifold's normal-bundle datum is all the extra information.","key_machinery":"The paper's load-bearing notion is strong multiplicativity: a $B$-structure with the property that, among a bundle $\\xi:E\\to M$, the tangent bundle of $E$, and the tangent bundle of $M$, any two $B$-structures force a canonical $B$-structure on the third. This is what makes the normal bundle of a submanifold in a pair a $B$-bundle, and it is what allows the kernel-cokernel computation in the split exact sequence. The geometric construction doing the work is the sphere bundle of $E\\oplus\\mathbb{R}$, where $E$ is the pullback of the universal bundle classified by a given normal-bundle datum; its 'north pole' section realizes any normal-bundle class and gives a way to replace a pair by a disjoint union of a pair with empty submanifold and a sphere-bundle pair, which is exactly the cancellation step in the proof. In the oriented part, the additional machinery consists of the bounded book bordism groups and the asymmetric signature as the unique obstruction for extending a book decomposition (for $n\\ge 5$), and the normal 1-type plus stable-diffeomorphism classification of 4-manifolds (for $n=4$).","core_discovery":"The central claim is Theorem 3.1: for a strongly multiplicative structure $B\\to BO_k$ with $k\\ge m>n$, the group of pairs $\\operatorname{SK}^B_{m,n}$ fits into a split short exact sequence $$0\\to \\operatorname{SK}^B_m\\to \\operatorname{SK}^B_{m,n}\\to \\operatorname{SK}^B_n(B_{m-n})\\to 0.$$ The inclusion sends a manifold to the same manifold with empty submanifold, and the surjection sends a pair $(M,N)$ to the submanifold $N$ together with the $B$-structure on its normal bundle in $M$, regarded as a map to the classifying space $B_{m-n}$. Specializing to the unoriented and oriented structures gives the split sequences for pairs with maps to $X$, and the paper's second main theorem shows that when $Y$ is simply connected and $\\pi_1(X)$ is finitely presented, projection induces an isomorphism $\\operatorname{SK}_n(X\\times Y)\\to \\operatorname{SK}_n(X)$. The proof strategy is to transplant the known unoriented pair argument into the $B$-setting using the sphere bundle of the normal bundle, and then to pass from maps to $X$ to maps to a product by proving a product formula for unoriented bordism and using algebraic surgery obstructions in high dimensions together with modified surgery in dimension 4.","pith_inferences":["The paper asserts without proof that $B\\times X$ is strongly multiplicative for any pointed path-connected $X$; this is exactly the kind of claim that should be checked against $X=S^1$, where the required two-out-of-three square either holds or fails by an explicit map. If it fails, the two splitting theorems for maps to $X$ could likely be repaired by restricting to spaces $X$ with a suitable con","The same split-sequence mechanism suggests that any tangential structure with a genuine two-out-of-three property, not just the examples listed, yields a pair-SK splitting; a categorical formulation of strong multiplicativity might make the proof independent of the ambient category.","Theorem 6.1's reliance on stable diffeomorphism in dimension 4 suggests a testable consequence: representatives of the zero class in $\\operatorname{SK}_4(X\\times Y)$ should be stably diffeomorphic to representatives coming from $X$ alone, so the difference between the two groups is controlled by stable-diffeomorphism invariants."],"forward_implications":["If $B$ is strongly multiplicative, then every computation of $\\operatorname{SK}^B_m$ and $\\operatorname{SK}^B_n(B_{m-n})$ yields a computation of the pair group $\\operatorname{SK}^B_{m,n}$ by a split extension.","In the unoriented case, $\\operatorname{SK}^O_{m,n}(X)$ is an extension of $\\operatorname{SK}^O_n(X\\times X)$ by $\\operatorname{SK}^O_m(X)$; the analogous statement holds with orientations, replacing $\\operatorname{SK}^O$ by $\\operatorname{SK}$.","The isomorphism $\\operatorname{SK}_n(X\\times Y)\\cong\\operatorname{SK}_n(X)$ for simply connected $Y$ means that adding a simply connected target component never creates new SK-obstructions.","The earlier triviality result for SK with a map into a simply connected space is recovered as the special case $X=\\mathrm{pt}$.","For pairs, the quotient term records the submanifold together with its normal-bundle structure, so pair invariants can distinguish submanifolds by that normal datum."],"supporting_citations":[{"why":"defines SK-groups and supplies the foundational exact sequences used in Section 4.","marker":"[KKNO73]"},{"why":"gives the unoriented cutting-and-pasting-of-pairs theorem whose proof Theorem 3.1 adapts to B-structures.","marker":"[Kom86]"},{"why":"proves that SK with a map into a simply connected space is trivial, the statement Theorem 6.1 generalizes, and supplies the dimension-2 computation.","marker":"[Neu75]"},{"why":"identifies unoriented bordism with homology over Z/2, the basis for the product formula used in Proposition 5.1.","marker":"[CF62]"},{"why":"provides the open-book-decomposition result used to show that manifolds with an open book decomposition vanish in SK.","marker":"[Win73]"},{"why":"supplies bounded book bordism groups and the asymmetric signature as the unique obstruction, used in the high-dimensional part of Theorem 6.1.","marker":"[Ran13]"},{"why":"gives the modified-surgery stable diffeomorphism classification used in the dimension-4 part of Theorem 6.1.","marker":"[Kre99]"},{"why":"states the normal 1-type of totally nonspin 4-manifolds with finitely presented fundamental group, used in the dimension-4 argument.","marker":"[KLPT17]"}],"fun_headline_variants":["SK of pairs splits via normal bundles for tangential structures","New split exact sequence for SK of manifold pairs","Tangential structures yield splitting of SK of pairs","SK of pairs splits via submanifold normal bundles"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The pair-splitting theorems for manifolds with a map to a space $X$ rest on the unproved step that adding an arbitrary pointed path-connected space $X$ to a strongly multiplicative tangential structure preserves the two-out-of-three rule; if this step fails, the normal bundle of the submanifold may lack a canonical structure and the proofs of Theorems 3.2 and 3.3 do not go through.","fun_headline_variants_meta":{"raw":{"variants":["SK of pairs splits via normal bundles for tangential structures","New split exact sequence for SK of manifold pairs","Tangential structures yield splitting of SK of pairs","SK of pairs splits via submanifold normal bundles"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000751,"raw_usage":{"total_tokens":3352,"prompt_tokens":961,"completion_tokens":2391,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":577,"completion_tokens_details":{"reasoning_tokens":2329}},"tokens_in":577,"tokens_out":2391,"duration_ms":18058,"temperature":1.0,"reasoning_tokens":2329,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T19:45:31.541590+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Check Example 2.6(3) for $X=S^1$, $r=s=1$. The pullback condition in the definition of strong multiplicativity requires an explicit comparison map out of $(BO_1\\times S^1)\\times(BO_1\\times S^1)$ to be a homeomorphism onto the fibre product determined by the bottom maps; writing both spaces as homotopy equivalent CW complexes and comparing $\\pi_1$ (or $\\pi_2$) is a finite check. If the square is not a pullback, the inheritance claim used for the normal-bundle structure of the submanifold is false; if it is, the claim survives this test.","supporting_citations":[],"review_version":1}