{"id":"9fcb92f4-7f4b-4af3-9f90-5962a86038c0","arxiv_id":"1908.09193","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"For each 1 <= p < 2, there exist orientation-preserving and orientation-reversing wild involutions of S^3 in the Sobolev class W^{1,p}(S^3,S^3).","lead":"The paper constructs wild involutions of the 3-sphere, self-inverse maps with a wild fixed-point set, that belong to the Sobolev space W^{1,p} for every 1 <= p < 2. This is the first Sobolev-regular wild involution construction and it connects decomposition topology to nonlinear elasticity.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Lemma 3.2's lift alignment condition has no integer solution for the one-cube rotations used in §4.2.3, leaving the central estimate (4.1) unsupported.","rationale":"The reader's weakest assumption points to the external Bing shrinkability input, but that input is classical and likely sound. The more fragile step is internal: Lemma 3.2 is the engine that converts Bing's twist foldings into the bilipschitz bound on h_ρ∘R±∘h_ρ^{-1}, and every L^p estimate in Section 5 flows from (4.1). The proof of Lemma 3.2 contains a concrete obstruction: the lift of the identity on the outer boundary and the lift of a nontrivial rotation on the inner loop have translation amounts that are incommensurate in the model case used by the construction. This is not a missing detail; the stated alignment condition cannot be satisfied for one-cube rotations. Because the theorem's construction requires such rotations in §4.2.3, the central claim is not established by the manuscript as written. The numerically false inequality in Section 5 (4^{k+1}(10L_inner^p)^{-k} ≤ (2L_inner^p)^{-k}) is a real error but repairable by enlarging the constant; it is not the primary obstruction. I therefore keep the reader's conditional verdict, but shift the condition away from the external Bing input and onto Lemma 3.2, which needs either a corrected lift construction or a different extension argument before Theorem 1.1 is supported.","tokens_in":44422,"tokens_out":21948,"duration_ms":236712,"concrete_test":"Verify the lift-alignment equation in Lemma 3.2 for the simplest case used in §4.2.2: take L0 a model loop with #L0=12, L'0 nested with side length σ/3 and #L'0=36, and let ρ be the one-cube shift. Write lifts explicitly: every lift of id on ∂|L| has form t↦t+12nσ; every lift of the one-cube rotation on |L'| has form t↦t+σ/3+12mσ on the boundary. The equality S(ι(x0))=~h'_ρ(S(x0)) requires 12nσ=σ/3+12mσ, i.e., 36n=1+36m, which has no integer solution. If this computation is correct, the construction in Lemma 3.2 cannot produce h_ρ for a one-cube shift, and the estimate (4.1) is unsupported; if the authors intended a different lift convention, they should specify it and recheck the deck-transformation claim.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Accepting the classical Bing/Montgomery–Zippin input, the new load-bearing estimate is Lemma 3.2: it supplies the twist h_ρ with Lip(h_ρ∘R±∘h_ρ^{-1}) ≤ L3 σ(L)/σ(L') (#L), which via (4.1) gives every L^p bound in §5. The proof of Lemma 3.2 constructs the lifted map ~h_ρ by affine interpolation and then asserts that ~h_ρ is a deck transformation of g_φ. That assertion is not verified, and the preceding lift alignment condition appears impossible for the rotations actually needed. Indeed, ~h'_ρ is a lift of a rotation by k_ρ cubes of the inner loop: in t it translates by k_ρσ(L') modulo the inner period #L'σ(L'); the boundary lift ι of the identity translates by a multiple of the outer period #Lσ(L). The condition S(ι(x0))=~h'_ρ(S(x0)) forces n#Lσ(L)=k_ρσ(L')+m#L'σ(L'). In §4.2.2, L' has side σ/3 and the same shape, so #L'=3#L and σ'=σ/3; the equation becomes k_ρ=3#L(n-m). For any nontrivial rotation, e.g., the one-cube shift used to realize Bing's foldings, k_ρ is between 1 and #L'-1, so the required divisibility by 3#L fails. Hence no such lift ι exists. The prescribed construction in Lemma 3.2 therefore does not produce h_ρ for the nontrivial rotations used in §4.2.3, and the bilipschitz estimate (4.1) has no basis.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper claims that for every 1 ≤ p < 2 there exist both an orientation-preserving and an orientation-reversing wild involution of S^3 in the Sobolev space W^{1,p}(S^3,S^3). The strategy is to realize Bing's wild involution as a quotient of a linear involution by a monotone map φ associated with Bing's double, and to choose a defining sequence of solid tori so that the composed map has local Lipschitz constants whose L^p norms can be controlled by selecting side lengths r_k. The technical apparatus consists of cubical loops and arcs, a bilipschitz straightening lemma for cubical arcs, and a lemma on twist maps of nested cubical loops; these feed into a local Lipschitz bound for the approximating involutions, followed by an L^p summability argument in Section 5.","tokens_in":44787,"tokens_out":14572,"duration_ms":146593,"significance":"If the proof can be repaired, the result is significant: it would show that wild topological involutions can have Sobolev regularity for every exponent below 2, complementing classical tameness results for C^1 involutions and the open problem of wild quasiconformal reflections. The constructive cubical approach, with explicit derivative estimates, is well motivated and connects to the Ball–Evans approximation problem via cellular mappings. However, the manuscript currently contains a serious gap in the proof of the central twist-map lemma and a separate incorrect estimate in Section 5, so the main theorem is not yet established by the arguments presented.","major_comments":[{"comment":"The prescribed construction of the ρ-twist h_ρ in Lemma 3.2 is invalid for the nontrivial rotations actually used. The lift ι of the identity on ∂|L| must satisfy S(ι(x0)) = ~h'_ρ(S(x0)); in the t-coordinate of the universal cover this forces n#Lσ(L) = k_ρσ(L') + m#L'σ(L'). In the nested pairs of §4.2.2 one has σ(L') = σ(L)/3 and #L' = 3#L, so the congruence reduces to k_ρ = 3#L(n−m). Hence no such lift exists for any rotation with 1 ≤ k_ρ ≤ #L'−1, in particular for the one-cube shift maps used in §4.2.3. Consequently ~h_ρ is not a deck transformation of the covering map g_φ, the map h_ρ does not descend, and the estimate (3.1) feeding into (4.1) is unsupported. The proof as written therefore does not establish Lemma 3.2.","section":"§3.2, Lemma 3.2"},{"comment":"The displayed inequality near the end of §5, namely ∫_{x∈X_k\\X_{k+1}: η(x)>0} |Df|^p ≤ (L_inner^k)^p (2L_inner^p)^{-k} = 2^{-k}, omits the factor 1/r_k^p coming from the local bound L(x) = L1 L_inner^{η(x)}/r_{ℓ(x)}. With r_k = (10L_inner^p)^{-k}, the missing factor contributes 10^{kp}L_inner^{kp^2}, which destroys the claimed summability. The estimate of the L^p norm on the corner set therefore needs to be redone, and the summability argument in terms of r_k must be corrected.","section":"§5, L^p estimate for the corner set"},{"comment":"The quantitative properties of the defining sequence—side-length ratios, nesting distances, symmetry, linking, and the identity #L' = 3#L—are asserted by reference to [5,7,25] rather than proved in the paper. These properties are load-bearing: they are used in the arithmetic of Lemma 3.2 and in the measure and volume estimates of §5. The authors should either provide a self-contained construction of such a defining sequence or give precise references with lemma and page numbers for each required property.","section":"§4.2.2, defining sequence"}],"minor_comments":[{"comment":"The abstract contains 'orientation-reserving' which should be 'orientation-reversing'.","section":"Abstract"},{"comment":"There is a typo 'collactions' for 'collections'.","section":"§2.1, Definition 2.4"},{"comment":"The index ℓ(x) in the local Lipschitz bound Lip f_k ≤ L1 L2^{η(x)}/r_{ℓ(x)} is never defined; it should be specified, e.g., as the level of the defining-sequence cube containing x.","section":"§4.2.4, Eq. (4.1)"},{"comment":"The notation for the constant in r_k = min{3^{-m_k}, 15^{-k}, (10L_p^2)^{-k}} is inconsistent with the rest of the paper; the third term should read (10L_inner^p)^{-k} to match the subsequent estimates.","section":"§5"}],"recommendation":"major_revision","confidential_remarks":"The two issues in Lemma 3.2 and Section 5 are genuine and load-bearing, but the overall approach is plausible and may be repairable. In particular, a direct construction of twist maps on the quotient torus might bypass the lift alignment obstruction, and the Section 5 estimate can likely be corrected with a more careful choice of r_k. I do not see grounds for outright rejection, but the paper needs substantial revision before the main theorem can be trusted."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Josh, quick take: this paper tackles a good question and the broad strategy is sensible, but the proof as written has a hole in its central lemma. The stress-test note is correct: Lemma 3.2's lift alignment condition forces the rotation step k_rho to be a multiple of #L', so no nontrivial rotation (including the one-cube shift used in §4.2.3) admits the claimed h_rho. That means the local bilipschitz estimate (4.1) has no basis, and the Lp estimates in §5 collapse unless the construction is changed.\n\nWhat is genuinely new and good: the idea of using uniformly bilipschitz cubical defining sequences to track derivative growth through Bing's shrinking process is a real step forward. The paper correctly identifies the heuristic threshold p=2, and the framing around nonlinear elasticity is well-motivated. The authors also give honest context about the open p>=2 question. If a correct version exists, it would be the first construction of Sobolev-regular wild involutions below p=2, which is a meaningful result.\n\nSoft spots: the displayed inequality in §5, as the reader noted, has a chain that fails for small k; that looks like a repairable arithmetic slip. The bigger problem is Lemma 3.2. The proof constructs a lift iota of the identity on the boundary satisfying S(iota(x0)) = h'_rho(S(x0)). Lifts of the identity are translations by multiples of #L sigma(L), while h'_rho translates by k_rho sigma(L') plus an inner period. In the configuration of §4.2.2, sigma(L') = sigma(L)/3 and #L' = 3#L, so the condition reduces to k_rho = 3#L(n-m). For any k_rho between 1 and #L'-1, no integers n,m exist. So the asserted lift does not exist. This is not a missing detail; it is a false assertion. Without it, the affine interpolation is not a deck transformation, and h_rho is not well-defined.\n\nWho is this for? Geometric topologists working on decomposition spaces and people in Sobolev homeomorphism theory. The theorem is worth knowing, but the current write-up is not ready. I would not cite it in this form.\n\nRecommendation: send it to a serious referee, because the question is important and the approach is promising. The referee should focus on Lemma 3.2 and ask whether a different extension trick can avoid the divisibility obstruction. That might be possible, for instance by not requiring the inner-line lift to be a translation by exactly k_rho cubes, but the current version does not do it.","headline":"Main result likely true but Lemma 3.2's construction fails for nontrivial rotations, so the proof needs major repair before the theorem is established.","tokens_in":45325,"tokens_out":5347,"would_cite":false,"duration_ms":49408,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["57S25","57R12","57N45","46E35","30C65"],"pacs":[],"model":"deepseek-v4-flash","headline":"For every $p<2$, there is a wild involution of the 3-sphere in $W^{1,p}$.","keywords":["wild involution","Sobolev homeomorphism","fixed point set","Bing double","monotone map","cubical loops","twist maps","decomposition space"],"falsifier":"A concrete way to test the claim is to check whether the defining sequence really supports the volume and corner-count bounds: if any symmetric shrinkable sequence forces the Lebesgue measure $|X_k|$ of the $k$-th stage to exceed $C\\,2^k r_k^2$ for a constant $C$ independent of $k$, then the integral estimate $\\int_{Y_k}|Df|^p \\lesssim 2^k r_k^{2-p}$ fails and the $p<2$ range would shrink. Alternatively, computing the actual $L^p$ norm of $|Df|$ for the explicit $r_k$ sequence on a computer model of the first few levels would give a numerical check of the summability.","tokens_in":44220,"feed_emoji":"🌀","tokens_out":7854,"duration_ms":73921,"temperature":0.7,"pith_summary":"The paper proves that wild topological involutions of the 3-sphere, whose fixed point sets are wildly embedded spheres or circles, can still be Sobolev regular: for every $1\\le p<2$ there is both an orientation-preserving and an orientation-reversing wild involution in the Sobolev space $W^{1,p}(S^3,S^3)$. This matters because it shows that topological wildness is compatible with quantitative analytic regularity just below the $W^{1,2}$ threshold. The construction is a Sobolev-space version of the classical Bing double construction, built from a monotone collapse of nested solid tori, and it supplies the same regularity for the monotone collapse map. The paper also leaves open whether the range $p\\ge 2$ can be reached and connects the question to approximation problems in nonlinear elasticity.","feed_headline":"Wild involutions of S3 exist at every Sobolev exponent below 2","feed_subtitle":"Bing's nested tori can be tuned so the involution's derivative is integrable for all exponents below 2.","key_machinery":"The load-bearing machinery is a cubical version of Bing's defining sequence: a tree of nested cubical loops, meaning solid tori made from dyadic cubes, with a symmetry plane at each corner. Inside each loop, a smaller nested loop is split into two arcs, and two linked loops are properly embedded; a $\\rho$-twist map $h_\\rho$ on a nested pair is a self-homeomorphism that is the identity on the outer boundary and realizes a rotation of the inner loop. Lemma 3.2 shows that every symmetric uniform nested pair admits such a twist whose conjugate involution $h_\\rho R h_\\rho^{-1}$ is locally bilipschitz with constant of order $(\\sigma(L)/\\sigma(L'))\\#L$, and Lemma 2.6 straightens cubical arcs bilipschitzly to segments. These estimates feed into a measure count: the corner set at level $k$ has controlled measure, and the non-corner part of $X_k$ has volume at most $2^{k+1}r_k^2$, so the chosen $r_k$ makes $|Df|^p$ summable.","core_discovery":"The central claim is Theorem 1.1: for each $p\\in [1,2)$, the 3-sphere admits an orientation-preserving and an orientation-reversing wild involution $f$ in $W^{1,p}(S^3,S^3)$. The proof constructs $f$ as the quotient of a linear involution by a monotone map $\\varphi$ that collapses the Bing double, a Cantor-set decomposition of $\\mathbb{R}^3$ whose nondegenerate elements are components of nested solid tori. A carefully chosen defining sequence of cubical tori with side lengths $r_k = \\min\\{3^{-k^{2/(2-p)}}, 15^{-k}, (10L_{\\mathrm{inner}}^p)^{-k}\\}$ makes $\\varphi$ and $f$ locally bilipschitz with constants whose singularities are integrable at exponent $p$; the contribution from corners is controlled by the $4^k$ count of corners, and the contribution from the remaining volume by the shrink rate. The regularity of $f$ is not an automatic consequence of the regularity of $\\varphi$, since $\\varphi\\circ\\iota\\circ\\varphi^{-1}$ need not lie in $W^{1,1}$; the twist structure supplies the additional cancellation.","pith_inferences":["A natural testable extension, not claimed in the paper, is that the same corner-count versus volume balance may produce wild $W^{1,p}$ involutions of $S^n$ in higher dimensions with a dimension-dependent threshold below $2$.","The existence of Sobolev-regular monotone cellular maps of this kind is a step toward a Sobolev version of cellular approximation, which would bear on the Lavrentiev phenomenon in three-dimensional elasticity.","The heuristic that $p=2$ is the critical balance between solid-torus volume and local bilipschitz rotation constant suggests that $p<2$ may be the sharp range for wildness; checking whether the adjoint integrability formula forces better regularity of $f^{-1}$ could test this."],"forward_implications":["The same proof yields a monotone map $\\varphi: S^3\\to S^3$ collapsing the Bing double in $W^{1,p}(S^3,S^3)$ for every $p<2$.","Wild fixed point sets, both a wildly embedded 2-sphere and a wildly embedded 1-sphere, are compatible with Sobolev differentiability of the involution for all $p<2$.","The composition $\\varphi\\circ\\iota\\circ\\varphi^{-1}$ is not automatically Sobolev from the Sobolev regularity of $\\varphi$; the twist cancellation built into the construction is essential.","The method reaches every $p<2$ but does not settle $p\\ge 2$; whether wild involutions exist in $W^{1,2}$ is left open."],"supporting_citations":[{"why":"Bing's original construction of an orientation-reversing wild involution and the Bing double decomposition that the paper makes Sobolev.","marker":"[5]"},{"why":"Bing's shrinking-without-lengthening argument supplies the foldings into twist maps and the corner counts used in the $L^p$ estimates.","marker":"[7]"},{"why":"Montgomery and Zippin's modification gives the orientation-preserving wild involution and its symmetry.","marker":"[25]"},{"why":"The technique of defining sequences of tori uniformly bilipschitz to model cubical tori in their inner metric is adapted from this work.","marker":"[15]"},{"why":"The Heinonen--Semmes question about wild quasiconformal reflections motivates asking how regular a wild involution can be.","marker":"[19]"},{"why":"Smith's fixed-point theorem for involutions of $S^3$ is the topological background that makes the wild fixed point set the object of interest.","marker":"[31]"},{"why":"Bochner's theorem that $C^1$ involutions have smooth fixed point sets frames the contrast the paper targets.","marker":"[8]"}],"fun_headline_variants":["Wild S3 involutions exist for every p<2","Sobolev p<2: wild involutions on S3","Bing's construction hits W^{1,p} for all p<2","Every p<2: a wild involution in Sobolev class"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The construction assumes that the Bing double has a defining sequence of nested solid tori that is simultaneously symmetric under the given involution, shrinkable in Bing's sense, and satisfies the explicit nesting distances and side-length bounds used in Section 5; this shrinkable defining sequence is taken from Bing's and Montgomery--Zippin's work rather than proved here.","fun_headline_variants_meta":{"raw":{"variants":["Wild S3 involutions exist for every p<2","Sobolev p<2: wild involutions on S3","Bing's construction hits W^{1,p} for all p<2","Every p<2: a wild involution in Sobolev class"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000216,"raw_usage":{"total_tokens":1375,"prompt_tokens":829,"completion_tokens":546,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":445,"completion_tokens_details":{"reasoning_tokens":469}},"tokens_in":445,"tokens_out":546,"duration_ms":5433,"temperature":1.0,"reasoning_tokens":469,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T11:19:21.166691+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A concrete way to test the claim is to check whether the defining sequence really supports the volume and corner-count bounds: if any symmetric shrinkable sequence forces the Lebesgue measure $|X_k|$ of the $k$-th stage to exceed $C\\,2^k r_k^2$ for a constant $C$ independent of $k$, then the integral estimate $\\int_{Y_k}|Df|^p \\lesssim 2^k r_k^{2-p}$ fails and the $p<2$ range would shrink. Alternatively, computing the actual $L^p$ norm of $|Df|$ for the explicit $r_k$ sequence on a computer model of the first few levels would give a numerical check of the summability.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Bing's original construction of an orientation-reversing wild involution and the Bing double decomposition that the paper makes Sobolev."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Bing's shrinking-without-lengthening argument supplies the foldings into twist maps and the corner counts used in the $L^p$ estimates."},{"cited_title":"Montgomery and L","cited_arxiv_id":null,"evidence_quote":"Montgomery and Zippin's modification gives the orientation-preserving wild involution and its symmetry."},{"cited_title":"Drasin and P","cited_arxiv_id":null,"evidence_quote":"The technique of defining sequences of tori uniformly bilipschitz to model cubical tori in their inner metric is adapted from this work."},{"cited_title":"Heinonen and S","cited_arxiv_id":null,"evidence_quote":"The Heinonen--Semmes question about wild quasiconformal reflections motivates asking how regular a wild involution can be."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Smith's fixed-point theorem for involutions of $S^3$ is the topological background that makes the wild fixed point set the object of interest."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Bochner's theorem that $C^1$ involutions have smooth fixed point sets frames the contrast the paper targets."}],"review_version":1}