{"id":"d47c7e66-b94d-4a2e-9815-590050e45b70","arxiv_id":"2506.11905","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A topological version of the Hatcher-Wagoner pseudo-isotopy obstructions is defined in dimension four and used to construct homeomorphisms of Y times S1 that are pseudo-isotopic but not isotopic to the identity.","lead":"This paper constructs topological invariants that detect when a 4-dimensional pseudo-isotopy cannot be deformed to an isotopy, matching the smooth invariants of Hatcher and Wagoner. The authors use these invariants to build homeomorphisms of Y times S1 that are homotopic but not isotopic to the identity.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 3.4 (TOP BLR connectivity) is sketched, not proved; Lemma 3.2 and Definition 3.11 of ΣTOP and ΘTOP rest on it, so without a complete proof the invariants are not established.","rationale":"The reader's weakest assumption correctly identifies Theorem 3.4 as the most fragile input. I agree with that assessment: unlike the allowed one-parameter families, where the paper gives precise definitions and a proof that they produce pseudo-isotopies, Theorem 3.4 is an unproved theorem imported from a sketch. It is not merely a convenience: the inverse map 𝔦 in Notation 3.6 is a formal consequence of Lemma 3.2, and every subsequent statement about ΣTOP and ΘTOP is proved by diagram chases that assume 𝔦 is a well-defined isomorphism. The topological category introduces real issues that the PL proof does not face, for instance the need to pass from concordance straightening to pseudo-isotopy spaces and the use of topological transversality that was not available to Pedersen. The paper is honest about this in Remark 1.8, but the admission confirms the gap rather than closing it. The allowed one-parameter families of Definition 5.2 are a more minor concern: the paper explicitly disclaims a full topological Cerf theory and only needs these families to construct the candidates for realisation, and the subsequent comparison with the stable smooth realisation is a reasonable strategy. A secondary inconsistency is that Theorem 7.1 is stated without the k1(X)=0 condition used in its proof; however, the main application in Theorems 9.3 and 1.5 is restricted to k1(Y)=0, so this does not affect the headline application. The appropriate verdict is unchanged from the reader's CONDITIONAL: the central claim should be accepted only after a complete proof of Theorem 3.4 is supplied.","tokens_in":44573,"tokens_out":15011,"duration_ms":243920,"concrete_test":"Extract Pedersen's appendix to [BLR75] and determine its exact scope: does it prove the needed π0/π1-connectivity for TOP pseudo-isotopy spaces, or only for concordance spaces? Then write out the full translation of [BLR75, Theorem 3.1′] to TOP, specifying where Morlet's disjunction lemma is used and verifying the bound (3.1) for the model case V = W ∪ (4-handles), with W a neighbourhood of the 3-skeleton. If the TOP pseudo-isotopy inclusion is not at least 1-connected in that model case, Lemma 3.2 is false and Definition 3.11 collapses.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The load-bearing step is Theorem 3.4, a TOP version of the Burghelea–Lashof–Rothenberg connectivity theorem. Lemma 3.2 uses it to show that the inclusion of a neighbourhood N of the 3-handle skeleton of X×J^2 induces an isomorphism on π0 of topological pseudo-isotopy spaces, and Section 4.0.1 uses it again for surjectivity from the 2-handle skeleton. The inverse map 𝔦 in Notation 3.6, which enters Definition 3.11 of ΣTOP and ΘTOP, exists only because of this theorem. The paper does not give a complete proof. The paragraph before Theorem 3.4 says the PL proof can be adapted to TOP using Pedersen's concordance straightening and topological transversality, but it does not state or prove the TOP Morlet disjunction lemma, does not carry out the adaptation of [BLR75, Lemma b)], and the footnote observes that Pedersen did not have [KS77] transversality available. The DIFF/PL arguments use smoothing and PL-specific tools that do not automatically transfer; moreover, concordance straightening applies to concordance spaces, whereas Theorem 3.4 concerns pseudo-isotopy spaces, so an additional fibration or homotopy-equivalence step is needed. Since well-definedness, naturality, compatibility with the smooth invariants, and duality are all proved by chasing diagrams through the same 𝔦, a failure or even a one-unit change in the connectivity bound in (3.1) would make the invariants ill-defined or non-canonical. The authors are transparent about this in Remark 1.8, but Theorem 1.1 depends on it.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper defines topological analogues, denoted ΣTOP and ΘTOP, of the Hatcher–Wagoner pseudo-isotopy obstructions for compact topological 4-manifolds. The construction double-suspends a 4-manifold X to a 6-manifold, uses a claimed TOP version of the Burghelea–Lashof–Rothenberg connectivity theorem (Theorem 3.4) to pass to a neighbourhood of the 3-handle skeleton, smooths that neighbourhood, and then evaluates the classical smooth invariants. The authors prove compatibility of the topological invariants with the smooth ones, naturality under certain codimension-zero inclusions, and a duality formula, and they state full realization theorems (Theorems 6.1 and 7.1) using 'allowed one-parameter families of topological handle decompositions' together with Freedman–Quinn disc embedding and Cha–Kim stable surface smoothing. These results are applied to construct, for many 3-manifolds Y, homeomorphisms of Y×S1 that are pseudo-isotopic to the identity but not isotopic to the identity (Theorem 1.5).","tokens_in":44879,"tokens_out":5673,"duration_ms":69973,"significance":"If the main theorems are correct, this is a substantial advance: it provides the first systematic topological obstruction theory for pseudo-isotopy versus isotopy in dimension four, compatible with the smooth Hatcher–Wagoner invariants, and yields new examples of homeomorphisms that are homotopic but not isotopic to the identity. The paper is transparent about its main technical debts, especially in Remark 1.8, and it makes productive use of recent deep results of Singh and of Cha–Kim. There is no circularity in the realization argument: the target values are compared with independent stable smooth realization results rather than built into the definitions. The classification and examples for 3-manifold groups (including the elliptic case) are concrete and useful. However, the central construction rests on a theorem that is asserted with only a sketch, and the realization mechanism is explicitly ad hoc; these points are load-bearing for Theorems 1.1, 1.2, and 1.5.","major_comments":[{"comment":"Theorem 3.4 is the sole basis for Lemma 3.2 and for the inverse map 𝔦 in Notation 3.6, which in turn enter the definition of ΣTOP and ΘTOP in Definition 3.11 and the compatibility proof in Lemma 4.2. The paragraph preceding Theorem 3.4 gives only a sketch: it asserts that the PL proof can be adapted to TOP using Pedersen's concordance straightening and topological transversality, but it does not state or prove a TOP Morlet disjunction lemma, does not carry out the adaptation of [BLR75, Lemma b)], and footnote 1 says the authors 'believe' Pedersen's approach can be simplified. Moreover, concordance straightening concerns concordance spaces, while Theorem 3.4 concerns pseudo-isotopy spaces, so an additional fibration or homotopy-equivalence step is needed. Since the bound (3.1) with r=3 and k=0 is exactly what yields π1(γ)=0 in Lemma 3.2, any change in the constants would require a new argument. I consider this the main correctness risk of the paper; a complete proof, or a precise citation to a fully proved TOP version, is necessary before the invariants are established.","section":"Section 3, Theorem 3.4"},{"comment":"The allowed one-parameter families of topological handle decompositions are the only source of candidate pseudo-isotopies in Construction 6.2 and Construction 7.2, so Theorem 1.2, and hence Theorems 9.3 and 9.11, depend on them. The paper explicitly disclaims having a topological Cerf theory and calls Definition 5.2 'ad hoc'. What is missing is a precise equivalence relation on such families together with a proof that the invariants of the resulting pseudo-isotopy are unchanged under the choices made in the construction (birth times, slide times, handle embeddings, auxiliary isotopies). In particular, the proofs of Lemma 5.6 and Lemma 5.7 combine an Alexander trick with the Freedman–Quinn procedure, but the subsequent comparison with the smooth realization results in Lemmas 6.3 and 7.3 is only up to a stable isotopy and leaves several choices unaddressed. For the realization theorem to be fully proved, this part needs to be made into a rigorous construction with well-defined outputs.","section":"Section 5, Definition 5.2"},{"comment":"The proof of the duality formula is not complete. The diagram introduces a map K described as 'straightening concordances' and states that the top square commutes by definition of K, but K is not identified with the isomorphism of Lemma 3.2 or with the map 𝔦 of Notation 3.6, and the interaction of duality with the smoothing and forgetful maps is only asserted. The proof also invokes Lemma 8.5, whose proof is attributed to Hatcher and only sketched, for the two suspension steps. Since Proposition 8.3 underlies Lemma 9.1 and therefore Theorem 9.3, the duality formula needs a full proof rather than a diagram with unexplained maps.","section":"Section 8, Proposition 8.3"}],"minor_comments":[{"comment":"The statement reads 'If either ΣTOP([F])≠0 and ΘTOP([F])≠0 then F is not topologically isotopic to an isotopy'; logically this should be 'if either ΣTOP([F])≠0 or ΘTOP([F])≠0', since the proof only needs one non-zero invariant.","section":"Theorem 3.14"},{"comment":"There is a spelling inconsistency in Section 9.4: Lemma 9.10 is cited as [Igu21a, Lemma 5.1] in the text but labelled as [Igu21b, Lemma 5.1] in the statement and the reference list; the intended reference should be made consistent.","section":"References and text"},{"comment":"The text writes 'Steifel-Whitney classes' in Proposition 8.3; the correct spelling is 'Stiefel-Whitney classes'.","section":"Section 8"},{"comment":"The acknowledgements refer to the 'Max Plank Institute'; this should be 'Max Planck Institute'. In addition, some figure references (e.g., Figures 2 and 3) are not cited in the running text where the corresponding Cerf graphics are discussed.","section":"Section 1.4"}],"recommendation":"major_revision","confidential_remarks":"This is a serious and potentially important paper, and the authors are appropriately candid about the gaps. My recommendation of major revision is driven by the fact that Theorem 3.4 is load-bearing and currently unproved, and by the need for a more rigorous treatment of the allowed one-parameter families. If the authors can supply a complete proof of the TOP Burghelea–Lashof–Rothenberg statement (or cite a published proof), and can tighten the construction and comparison in Sections 5–7, I would be happy to reconsider. The paper fits the journal's scope well, and the application to Y×S1 is attractive, but the current manuscript should not be accepted before these points are resolved."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dan, this paper is real. It defines topological Hatcher-Wagoner invariants Sigma_{TOP} and Theta_{TOP} for 4-manifolds, proves they agree with the smooth invariants, and proves full realisation for good fundamental groups. That is genuinely new work, not a re-coordinatization of known results, and it corrects a real gap in Kwasik's old sketch. The construction of homeomorphisms of Y x S^1 that are pseudo-isotopic but not isotopic to the identity is a nice payoff.\n\nThe soft spots are exactly where the stress-test note puts them. Theorem 3.4, the TOP version of Burghelea-Lashof-Rothenberg connectivity, is load-bearing: Lemma 3.2 and the inverse map i in Definition 3.11 depend on it, and the proof is only a sketch. The paper says Pedersen's concordance straightening plus topological transversality adapts the PL proof, but it does not prove the TOP Morlet disjunction lemma or carry out the adaptation of BLR Lemma b). The footnote even concedes that Pedersen did not have Kirby-Siebenmann transversality available. For a theorem on which the main invariants rest, this is a real hole, though not obviously fatal.\n\nThe second issue is Definition 5.2's 'allowed one-parameter families'. They are explicitly ad hoc, and the authors admit there is no topological Cerf theory. The realisation theorem uses them to produce pseudo-isotopies, and the comparison with smooth stable realisation via Cha-Kim is clever. But it is fair to worry whether these families capture enough of the smooth theory to justify the claimed full realisation.\n\nI do not see a hidden circularity; the comparisons to Singh and Cha-Kim are independent. The paper is unusually honest about what is proved and what is sketched. If Theorem 3.4 can be fully established, the invariants stand. If not, the definition itself collapses. That is the correct question for a referee to press.\n\nThis paper deserves a serious referee, and I would send it to a good topology journal. The referee should demand a complete proof of Theorem 3.4 in the TOP category, or a precise reduction to a stated theorem, before publication. I would also ask for a clearer justification that the allowed families in Section 5 connect to the standard smooth one-parameter families used in the invariants.\n\nFor my own work, I would cite it as a conditional but significant advance, and I would bring it to the reading group.","headline":"Serious, honest paper with a load-bearing sketched TOP theorem; worth refereeing but not yet fully established.","tokens_in":45429,"tokens_out":2270,"would_cite":true,"duration_ms":29090,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["57K40","57R50","57R52","19B28"],"pacs":[],"model":"deepseek-v4-flash","headline":"Two Whitehead-valued obstructions separate pseudo-isotopy from isotopy in topological 4-manifolds, and yield non-isotopic homeomorphisms of Y×S1.","keywords":["topological pseudo-isotopy","4-manifolds","isotopy","Whitehead groups","smooth obstruction theory","disc embedding theorem","homeomorphism groups","good fundamental groups"],"falsifier":"Test the isomorphism of Lemma 3.2 on a compact topological 4-manifold with nontrivial $\\pi_2$: compare $\\pi_0$ of the TOP pseudo-isotopy space of the twice-suspended manifold with $\\pi_0$ of the space for a neighbourhood of its 3-handle skeleton, and check whether the inclusion is a bijection. A single example where the inclusion induces a nonzero relative homotopy class in degree 0 or 1 would make Definition 3.11 ill-defined and Theorem 1.1 collapse.","tokens_in":44300,"feed_emoji":"","tokens_out":13844,"duration_ms":144009,"temperature":0.7,"pith_summary":"In dimension four, a homeomorphism of a compact manifold can be pseudo-isotopic to the identity—linked to it by a homeomorphism of $X \\times I$ that is not required to preserve levels—without being isotopic to it. The paper defines two algebraic homomorphisms, $\\Sigma^{\\mathrm{TOP}}$ and $\\Theta^{\\mathrm{TOP}}$, valued in Whitehead groups built from $\\pi_1$ and $\\pi_2$ of the manifold, that obstruct a topological pseudo-isotopy from being a genuine isotopy. These match the classical smooth invariants when the manifold carries a smooth structure, making the construction a topological analogue in dimension four. The paper proves the invariants are fully realisable for manifolds with good fundamental group, and uses this to construct, for many closed 3-manifolds $Y$, homeomorphisms of $Y \\times S^1$ that are pseudo-isotopic and homotopic to the identity but not isotopic to it.","feed_headline":"Invariants catch pseudo-isotopies that are not isotopies","feed_subtitle":"For many 3-manifolds Y, the paper builds homeomorphisms of Y×S1 that are homotopic to the identity but not isotopic.","key_machinery":"The carrying mechanism is a chain of reductions. A pseudo-isotopy of $X$ is suspended twice to a pseudo-isotopy of the 6-manifold $X \\times J^2$; a topological analogue of the high-dimensional connectivity theorem from [BLR75] (Theorem 3.4) identifies $\\pi_0$ of the TOP pseudo-isotopy space with that of a neighbourhood $N$ of the 3-handle skeleton. Smoothing theory for topological manifolds turns $N$ into a smooth manifold, where the classical smooth invariants $\\Sigma$ and $\\Theta$, defined through one-parameter families of handle decompositions and their graphics tracking births, deaths, and handle slides, can be evaluated; a comparison with the 2-handle skeleton shows the final value does not depend on the chosen smoothing. For the realisation direction, the paper introduces 'allowed one-parameter families of topological handle decompositions', an explicit ad hoc stand-in for the missing topological Cerf theory, and uses the disc embedding theorem to put 2- and 3-handles into topological cancelling position so the families terminate in genuine pseudo-isotopies realising prescribed Whitehead elements.","core_discovery":"The central claim is that a two-stage obstruction theory, previously known for smooth pseudo-isotopies in high dimensions, exists for topological pseudo-isotopies of compact 4-manifolds. For any such $X$ the paper defines $\\Sigma^{\\mathrm{TOP}} : \\pi_0 \\mathcal{P}^{\\mathrm{TOP}}(X,\\partial X) \\to Wh_2(\\pi_1 X)$ and $\\Theta^{\\mathrm{TOP}} : \\ker \\Sigma^{\\mathrm{TOP}} \\to Wh_1(\\pi_1 X; \\mathbb{Z}/2 \\times \\pi_2 X)/\\chi$, and proves both vanish on pseudo-isotopies that are topologically isotopic to isotopies and agree with the smooth invariants under the forgetful map. When $\\pi_1 X$ is good, Theorem 1.2 asserts every element of either Whitehead group is realised by an actual pseudo-isotopy. Combining realisation with a duality formula for inertial pseudo-isotopies gives Theorem 1.5: if $Y$ is a closed 3-manifold with trivial first $k$-invariant and good, non-ambivalent fundamental group, then $Y \\times S^1$ has a homeomorphism that is pseudo-isotopic to the identity, homotopic to the identity, and not isotopic to the identity.","pith_inferences":["Going beyond the paper, if the invariants are well defined, the quotient by inertial pseudo-isotopies should give a well-defined invariant on mapping classes of 4-manifold homeomorphisms, offering the first systematic algebraic way to decide pseudo-isotopy versus isotopy in the topological category.","The ad hoc 'allowed one-parameter families' are a placeholder for the missing topological Cerf theory; the realisation results suggest that any future topological Cerf theory will be compatible with these invariants on the constructed pseudo-isotopies.","The non-ambivalence condition is likely the right general obstruction to the inertial subgroup: for any group with a conjugacy class not equal to its inverse, the same construction should yield non-isotopic pseudo-isotopic homeomorphisms once the $k$-invariant and goodness hypotheses are met, so the examples listed may be the tip of a larger class.","A natural stress test is to compute $\\Theta^{\\mathrm{TOP}}$ on smooth pseudo-isotopies previously constructed by other methods in dimension four; if they survive suspension, the topological and smooth invariants would agree there too."],"forward_implications":["If Theorem 1.1 holds, every smooth pseudo-isotopy detected by the classical two-stage invariants in a smooth 4-manifold is also detected by $\\Sigma^{\\mathrm{TOP}}$ and $\\Theta^{\\mathrm{TOP}}$, so the topological theory is at least as strong as the smooth theory where both are defined.","Theorem 1.2 gives full unstable realisation for good fundamental groups: every class in $Wh_2(\\pi_1 X)$ and in $Wh_1(\\pi_1 X; \\mathbb{Z}/2 \\times \\pi_2 X)/\\chi$ arises as $\\Sigma^{\\mathrm{TOP}}$ or $\\Theta^{\\mathrm{TOP}}$ of some topological pseudo-isotopy.","Theorem 1.5 produces many closed 4-manifolds of the form $Y \\times S^1$, including lens spaces, tetrahedral manifolds, most icosahedral manifolds, and the 3-torus, carrying homeomorphisms that are homotopic to the identity and pseudo-isotopic to it but not isotopic to it.","The naturality and duality formulas give computable control of the inertial pseudo-isotopy subgroup, making the obstruction to a homeomorphism well defined modulo this subgroup and allowing the distinction to pass from $Y \\times I$ to the closed manifold $Y \\times S^1$."],"supporting_citations":[{"why":"Supplies the smooth two-stage obstruction theory that the paper's topological invariants are designed to match.","marker":"[HW73]"},{"why":"Provides the high-dimensional pseudo-isotopy obstruction framework and the connectivity theorem whose topological analogue is used to reduce to the 3-handle skeleton.","marker":"[BLR75]"},{"why":"Gives the comparison between smooth and topological pseudo-isotopy spaces used to pass from the smoothed skeleton back to the topological category.","marker":"[BL74]"},{"why":"Supplies concordance straightening and concordance-implies-isotopy, needed for the topological connectivity theorem and the smoothing passage.","marker":"[Ped77]"},{"why":"Supplies the disc embedding theorem, the notion of good fundamental group, and handle-cancellation techniques used in the realisation construction.","marker":"[FQ90]"},{"why":"Gives the stable smooth realisation results for the two invariants in dimension four that the paper's unstable topological realisation is compared against and refines.","marker":"[Sin22]"},{"why":"Supplies the stable surface smoothing theorem used to compare the topological Whitney-disc construction with smooth stable realisations.","marker":"[CK23]"},{"why":"Provides smoothing theory and topological transversality for the handle-skeleton smoothing step and the handle decomposition of the suspended manifold.","marker":"[KS77]"}],"fun_headline_variants":["4D obstructions separate pseudo-isotopy from isotopy","Topological invariants catch non-isotopic pseudo-isotopies","Obstructions that forbid pseudo-isotopies from being isotopies","Pseudo-isotopy vs isotopy: 4D obstacles"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"Everything rests on the unproved topological analogue of the high-dimensional connectivity theorem from [BLR75] (Theorem 3.4), which says that after two suspensions a pseudo-isotopy is determined up to isotopy by its behaviour near the 3-handle skeleton; the paper sketches the adaptation via concordance straightening and topological transversality but supplies no complete proof.","fun_headline_variants_meta":{"raw":{"variants":["4D obstructions separate pseudo-isotopy from isotopy","Topological invariants catch non-isotopic pseudo-isotopies","Obstructions that forbid pseudo-isotopies from being isotopies","Pseudo-isotopy vs isotopy: 4D obstacles"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000793,"raw_usage":{"total_tokens":3475,"prompt_tokens":908,"completion_tokens":2567,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":524,"completion_tokens_details":{"reasoning_tokens":2493}},"tokens_in":524,"tokens_out":2567,"duration_ms":22610,"temperature":1.0,"reasoning_tokens":2493,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T01:01:15.114149+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Test the isomorphism of Lemma 3.2 on a compact topological 4-manifold with nontrivial $\\pi_2$: compare $\\pi_0$ of the TOP pseudo-isotopy space of the twice-suspended manifold with $\\pi_0$ of the space for a neighbourhood of its 3-handle skeleton, and check whether the inclusion is a bijection. A single example where the inclusion induces a nonzero relative homotopy class in degree 0 or 1 would make Definition 3.11 ill-defined and Theorem 1.1 collapse.","supporting_citations":[],"review_version":1}