{"id":"da019741-06e7-4228-961f-7e3fe17bad58","arxiv_id":"2501.05489","paper_version":3,"verdict":"CONDITIONAL","confidence":"LOW","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For closed smooth mean curvature flows in R^{n+1}, 3≤n≤6, at the first singular time either the mean curvature or the Morse index must blow up.","lead":"This paper proves that a closed smooth mean curvature flow in dimensions 3 through 6 cannot form a singularity unless either its mean curvature or its Morse index goes to infinity. It also gives new estimates on the size of the singular set when both quantities stay bounded.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 2.6, the two-sided pseudolocality estimate for flows with bounded mean curvature, is stated without proof; the claimed Minkowski-dimension bound for the singular set depends on it, so the full Theorem 0.4 is not yet verified.","rationale":"The reader's weakest-assumption identification is correct: Theorem 2.6 is stated without proof and is used at a key point. My stress-test confirms that the proved pseudolocality theorem, Theorem 2.5, is not a substitute: it requires small mean curvature and, as written, its Ecker-Huisken step controls only forward times. The unproved two-sided version is needed for Lemma 6.5 and hence for the quantitative singular-set estimates and the Minkowski dimension part of Theorem 0.4. The multiplicity-one convergence argument in Sections 3-5 may be fillable, and the reader's CONDITIONAL verdict appropriately reflects that. I do not find a separate internal inconsistency that would force REJECT: the dimensional issue in Theorem 6.8's volume estimates appears to be resolved if 'Vol' is interpreted as the ambient (n+1)-dimensional tubular volume, which matches the claimed Minkowski dimension n−7. Therefore the verdict should remain CONDITIONAL, pending a proof of Theorem 2.6 or a clear derivation of Lemma 6.5 from the small-H theorem via scaling and a covering argument.","tokens_in":24893,"tokens_out":27886,"duration_ms":290776,"concrete_test":"Provide a complete proof of Theorem 2.6 by adapting the proof of Theorem 2.5 to the case |H| ≤ Λ. Concretely, track the evolution of the graphical radius in the average-normal estimate (2.14): compute the maximal time |t| for which the graph over a fixed plane remains controlled when the mean-curvature bound is Λ, and check whether it is at least η r0^2/(Λ+Λ^2) for some η with lim_{Λ→0} η = η0 > 0. If the maximal window is smaller, test Lemma 6.5 on a rescaled mean curvature flow with bounded H and a conical singularity to see whether r_M(X) ≥ r fails at some X in P_{β^{-1}r}(0). Either a full proof of Theorem 2.6 or one concrete counterexample would settle the concern.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The load-bearing gap is Theorem 2.6 in Section 2, which is introduced with 'We omit the proof here'. This two-sided pseudolocality estimate is precisely what Lemma 6.5 uses to convert a time-0 regularity-scale bound into a space-time regularity-scale bound, and Theorem 6.8 then uses Lemma 6.5 to obtain the quantitative volume estimates (6.11)-(6.12) and the Minkowski dimension bound dim_Mink(S) ≤ n−7 in Theorem 0.4. The proved Theorem 2.5 does not cover this use: it assumes mean curvature bounded by a small δ, and its proof applies the Ecker-Huisken interior estimate (Lemma 2.1) in a forward time direction, so it gives curvature control only for t ≥ 0 after the initial slice. Lemma 6.5 requires control on a full parabolic cylinder P_{β^{-1}r}(0), including negative times. Thus the stated proof of the main theorem has an unproved input at a central junction. If Theorem 2.6 fails in the stated form, the convergence to stable cones with multiplicity one might still hold, but it would not yield the Minkowski dimension conclusion for S or the parabolic estimates of Theorem 0.6. The n ≤ 6 no-blow-up statement might survive via the density-one argument, but the theorem as stated, including the singular-set dimension for n ≥ 7, is conditional on this estimate.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies closed smooth embedded mean curvature flows in R^{n+1} for n ≥ 3 under two global bounds: a uniform L∞ bound on the mean curvature and a uniform bound on the Morse index of the time slices. The main theorem (Theorem 0.4) claims that, under these bounds, the flow converges smoothly with multiplicity one to a limit hypersurface away from a singular set of Minkowski dimension at most n−7; in particular, for 3 ≤ n ≤ 6 the flow does not blow up at the first singular time. The proof follows the Li-Wang strategy: a two-sided pseudolocality estimate, weak compactness theorems for hypersurfaces and flows, rescaling to stable minimal cones, and a multiplicity-one argument based on L-stability. Theorem 0.6 states quantitative estimates for the space-time singular set. The paper also states a corollary for n = 7 with a discrete singular set.","tokens_in":25161,"tokens_out":20438,"duration_ms":212275,"significance":"If the results are fully established, they would be a substantial advance: the n ≤ 6 no-blow-up statement under bounded mean curvature and bounded index directly addresses a conjecture related to Ilmanen's multiplicity-one conjecture, and the quantitative singular-set estimates for n ≥ 7 are in the spirit of Cheeger-Haslhofer-Naber. The paper is carefully organized around standard compactness, regularity, and stability tools, and it explicitly credits prior work through precise citations. However, the central proof currently depends on an unproved two-sided pseudolocality theorem (Theorem 2.6), with the manuscript stating 'We omit the proof here.' As a result, the main theorems should be regarded as conditional until that input is supplied or replaced.","major_comments":[{"comment":"Theorem 2.6 is a load-bearing input for the main theorems, but it is stated without proof: the text after the statement reads 'We omit the proof here.' This estimate is used in Lemma 6.5 to pass from the time-zero regularity scale r_{M0}(0) to the space-time regularity scale r_M(X) on a full parabolic cylinder, and Theorem 6.8 then uses Lemma 6.5 to obtain the volume estimates (6.11)-(6.12) and the Minkowski dimension bound in Theorem 0.4. The proved Theorem 2.5 does not cover Theorem 2.6: its proof uses the smallness of δ in an essential way (for example in (2.14)) and applies the Ecker-Huisken interior estimate (Lemma 2.1) only for positive times, so it gives no backward-in-time curvature control. Please provide a complete proof of Theorem 2.6, a precise reference, or a clear statement of which theorems are conditional on it.","section":"Section 2, Theorem 2.6"},{"comment":"Lemma 5.7 defines the quantity s_C(x) only for x ∈ reg(C), but its conclusion is stated for all x ∈ (C ∩ B_R(0)) \\ H(C,ε,R). The set H(C,ε,R) is a neighborhood of the low-curvature set S, not of the singular set of C, so points of sing(C) are included in the stated domain of the conclusion. The proof asserts that because x_j ∉ H(C_j,ε,R), 'C_j smoothly converges to C near x∞'; this is unjustified when x∞ ∈ sing(C), and the class C(N,n) contains singular cones. Lemma 5.8 (|T_N(C,ε,ζ,R)| = 0) and Lemma 5.9 use the conclusion of Lemma 5.7 for the limiting cones, so the thin-part argument is incomplete unless the singular set is handled separately or shown to be negligible.","section":"Section 5, Lemma 5.7"},{"comment":"The proof of Lemma 5.3 asserts, without argument, property (d) of the renormalized sequence, namely a uniform lower bound on the area ratio; this lower bound is part of the definition of a refined sequence and is used in Proposition 4.2. Lemma 5.9 also asserts convergence of the thin parts |T_N(Σ_{t_i},ε,ζ,R)| to |T_N(Σ∞,ε,ζ,R)| even though the convergence in Lemma 5.3 is only smooth away from sing(Σ∞) ∪ S_0, and points approaching the singular set or the time-dependent set e^{t/2}S_0 are not controlled. Please supply the missing estimates or state explicitly how the singular and exceptional sets are bypassed in this convergence.","section":"Section 5, Lemmas 5.3 and 5.9"}],"minor_comments":[{"comment":"The regularity scale is defined by 'sup_{y∈M∩B_r(y)} r|A|(y) ≤ 1'; the ball should be centered at the point x where the scale is being evaluated, not at the running point y.","section":"Section 5, Definition 5.4"},{"comment":"The displayed inequality 'C_2(ε,S,T,x_0) < w_i(x,t) < C_1(ε,K,S_0,x_0) > 0' contains a typo; the intended statement is 0 < C_2 < w_i(x,t) < C_1.","section":"Section 5, equation (5.29)"},{"comment":"The statement of Theorem 2.6 begins 'For any r ∈ (0,1], T ≥ 1/2 and Λ > 0' but the hypotheses and conclusion use r_0; the notation should be aligned.","section":"Section 2, Theorem 2.6"},{"comment":"The lemma says 'There exists ζ_0(R,N,ρ) > 0 with that for any C ∈ C(N,ρ)', even though no parameter ρ has been introduced in the statement; this should read C(N,n) and the dependence of ζ_0 should be stated consistently.","section":"Section 5, Lemma 5.7"},{"comment":"The proof of the index-zero case of Theorem 3.3 is delegated to 'repeat the proof of Theorem 2 in [32]' with no detailed adaptation to the bounded-mean-curvature setting; since this theorem is central to the paper, please either include the adapted proof or state precisely which arguments from [32] carry over unchanged.","section":"Appendix A and Theorem 3.3"}],"recommendation":"major_revision","confidential_remarks":"The paper presents a serious and potentially significant approach, but the proof of the main theorem is conditional on the unproved Theorem 2.6, and there are additional gaps in the thin-part argument around singular points of the limit cone. I recommend major revision rather than rejection because these are missing proofs rather than identified false statements; if the author supplies a complete proof of Theorem 2.6 and clarifies Lemmas 5.7-5.9, the results would merit publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The result is substantial and, if completed, would resolve a conditional form of Ilmanen's multiplicity-one problem in dimensions 3 through 6. The main theorem — bounded |H| and bounded Morse index imply smooth multiplicity-one convergence away from a set of Minkowski dimension at most n−7, and hence the dichotomy for 3 ≤ n ≤ 6 — is genuinely new. The weak compactness theorems (3.3 and 4.2) are useful tools in their own right, and the extension of Li–Wang's height-difference argument to the bounded-mean-curvature setting is a real step forward, not a routine transcription. I also see no circular reasoning: the proof derives the dichotomy from established compactness, regularity, and stability theorems, plus new technical inputs.\n\nThe soft spot is exactly where the stress-test note puts it. Theorem 2.6 is introduced with \"We omit the proof here,\" and it is not a peripheral lemma. Lemma 6.5 uses Theorem 2.6 to convert a time-0 regularity-scale bound into a space-time regularity-scale bound on a full parabolic cylinder, and Theorem 6.8 then uses that to get the quantitative volume estimates and the Minkowski dimension bound in Theorem 0.4. Without a proof of Theorem 2.6, the singular-set dimension and Theorem 0.6 are conditional. The n ≤ 6 no-blow-up statement might survive through the density-one argument, but the paper as written does not verify the full theorem.\n\nThere are smaller gaps in the same direction. Lemma 5.7's compactness argument has an unstated case when the limit point lies on the singular set of the limit cone: varifold convergence does not by itself justify \"smoothly converges near x∞,\" so the contradiction is not fully established. Lemma 5.3 asserts the uniform lower area ratio needed to form a refined sequence, and Lemma 5.10 asserts without proof that the relevant components remain in a fixed tubular neighborhood of the limit cone. These look fillable, but they are not just cosmetic. I would also flag that Theorem 6.8 is labeled with reference [3], but the statement is for Brakke flows and depends on Lemma 6.5 and Theorem 6.7; the attribution is at best imprecise and should be cleaned up.\n\nMy guess is that Theorem 2.6 is true and can be proved by adapting the proof of Theorem 2.5 with a scale-dependent cutoff and the mean curvature bound, but that is a guess. As submitted, the paper is a promising proof sketch with one central unproved input and a few technical assertions that need checking. It deserves peer review, not desk rejection, but the referee should expect a major revision. I would not cite the main theorem until Theorem 2.6 is proved.","headline":"A genuinely new multiplicity-one dichotomy for MCF with bounded mean curvature and bounded index, but the version on arXiv is conditional: the load-bearing two-sided pseudolocality estimate, Theorem 2.6, is stated without proof.","tokens_in":25724,"tokens_out":2932,"would_cite":false,"duration_ms":32110,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["53E10","53A10","49Q20"],"pacs":[],"model":"deepseek-v4-flash","headline":"For closed embedded mean curvature flows in dimensions 3 through 6, a first singularity forces either the mean curvature or the Morse index to blow up; in higher dimensions the singular set has Minkowski dimension at most n−7.","keywords":["mean curvature flow","Morse index","multiplicity-one conjecture","singularities","minimal cones","pseudolocality","stable minimal hypersurfaces","Minkowski dimension"],"falsifier":"A concrete search: look for a smooth embedded mean curvature flow with uniformly bounded mean curvature whose time-zero slice has area ratio $1+o(1)$ in a ball, but whose second fundamental form exceeds $1/(\\varepsilon r_0)$ inside the parabolic cylinder predicted by Theorem 2.6. Finding one would disprove the unproved pseudolocality estimate and collapse the proof's foundation; proving Theorem 2.6 would close the paper's only explicit gap.","tokens_in":60,"feed_emoji":"","tokens_out":13838,"duration_ms":192443,"temperature":0.7,"pith_summary":"The paper proves that a closed smooth embedded mean curvature flow in $\\mathbb{R}^{n+1}$ with uniformly bounded mean curvature and uniformly bounded Morse index cannot produce a first singular time when $3 \\leq n \\leq 6$: it must converge smoothly through that time. In all dimensions $n \\geq 3$, the flow converges smoothly with multiplicity one to a limit hypersurface away from a singular set of Minkowski dimension at most $n-7$, and every blow-up limit is a stable minimal cone obtained with multiplicity one. The equivalent dichotomy is that at the first singular time, either the mean curvature or the Morse index (the number of negative eigenvalues of $\\Delta+|A|^2$) must blow up. The interest is that bounded index and bounded mean curvature together rule out the stacked, multiple-sheet singular behavior that has been the main obstacle in the multiplicity-one problem.","feed_headline":"Bounded mean curvature and index rule out blow-up in 3–6","feed_subtitle":"If both stay bounded up to time T, the flow stays smooth through T, so a singularity must make one diverge.","key_machinery":"The load-bearing mechanism is the two-sided pseudolocality estimate, Theorem 2.6: if a time slice is nearly Euclidean in a ball and the whole flow has uniformly bounded mean curvature, then the second fundamental form is controlled in a small spacetime cylinder. Around this, the paper builds a weak compactness theory for hypersurfaces with bounded mean curvature, area ratio, and Morse index (Theorem 3.3), which gives $C^{1,\\alpha}$ convergence away from a finite set of at most $I$ points where index concentrates, with $H^{n-2}(\\operatorname{sing} M)=0$. The multiplicity-one argument then passes to the rescaled flow and studies the height difference between the top and bottom sheets of a putative multi-sheeted limit over a stable cone: the normalized height difference satisfies a parabolic equation close to the linearized shrinker operator $L = \\Delta - \\frac{1}{2}\\langle x,\\nabla\\cdot\\rangle + |A|^2 + \\frac{1}{2}$, and parabolic Harnack estimates force a positive solution of $Lw=0$ on the cone. A positive solution of that sign would make the cone $L$-stable, which contradicts the strong Frankel theorem for shrinkers: no proper $F$-stationary rectifiable hypersurface with $H^{n-2}(\\operatorname{sing} V)=0$ is $L$-stable. Hence the multiplicity must be one.","core_discovery":"The paper establishes Theorem 0.4: for $n \\geq 3$, let $M_t$ be a closed smooth embedded mean curvature flow in $\\mathbb{R}^{n+1}$ on $[0,T)$ with $\\sup |H| = \\Lambda < \\infty$ and $\\sup_t \\operatorname{index}(M_t) = I < \\infty$. Then there is a limit hypersurface $\\bar M_T$ and a subset $S \\subset \\bar M_T$ such that $M_t$ converges smoothly to $\\bar M_T$ away from $S$ with multiplicity one, and $S$ has Minkowski dimension at most $n-7$. Consequently, for $3 \\leq n \\leq 6$, the flow does not blow up at time $T$. The proof also shows that for any point $p \\in \\bar M_T$ and any sequence $t_i \\to T$, the rescaled surfaces $\\frac{1}{\\sqrt{T-t_i}}(M_{t_i}-p)$ converge, after subsequence, to a stable minimal cone with multiplicity one; in low dimensions the only such cone is a plane, so every point has Gaussian density one and is regular by the standard local regularity theorem for mean curvature flow.","pith_inferences":["The paper leaves Theorem 2.6, the two-sided pseudolocality estimate for bounded mean curvature, explicitly unproved; a reader should treat the theorem as conditional on that estimate, since all later compactness steps use it.","The codimension-7 bound matches the singular-set dimension for stable minimal hypersurfaces, suggesting that bounded index upgrades the singular behavior of a bounded-mean-curvature flow to the stable-minimal setting; sharpness could be probed with known non-flat stable minimal cones in dimensions $n \\geq 7$.","The dichotomy raises a quantitative question the paper does not answer: if $|H|$ stays bounded at a singularity in dimensions 3 through 6, how fast must the Morse index grow, and is the growth rate determined by the singularity type?","The same height-difference mechanism might extend to flows with controlled index growth or to ambient manifolds with positive Ricci curvature, where the final Frankel-type obstruction would need a different form."],"forward_implications":["For $3 \\leq n \\leq 6$, any closed smooth embedded mean curvature flow that reaches a first singular time must have either unbounded mean curvature or unbounded Morse index; if both stay bounded, the flow remains smooth through that time.","For $n \\geq 7$, under the same bounds the flow converges smoothly with multiplicity one to a limit hypersurface away from a singular set of Minkowski dimension at most $n-7$, and every blow-up limit is a stable minimal cone with multiplicity one.","In dimension 7 the singular set is discrete, and near each singular point the rescaled flow approaches a stable regular cone.","Quantitatively, the set where the parabolic regularity scale is below $r$ has volume at most $C(1+I)r^8$ in spacetime and $C(1+I)r^7$ at each time slice, giving finite $(n-7)$-dimensional Hausdorff measure for the time-slice singular set and finite $(n-5)$-dimensional measure for the spacetime singular set.","The theorem gives a clean dichotomy at the first singular time: either the mean curvature or the Morse index blows up, with no third possibility under the smooth embedded assumption."],"supporting_citations":[{"why":"Provides the Gaussian monotonicity formula, density bounds, and the distance estimate used at the final rescaling to show regularity.","marker":"[21]"},{"why":"Supplies the pseudolocality lemmas, the decomposition of space into high-curvature, thick, and thin parts, and the overall multiplicity-one strategy.","marker":"[26]"},{"why":"Supplies the height-difference parabolic equation, the parabolic Harnack estimates, and the extraction of a positive solution on the limit cone.","marker":"[27]"},{"why":"Provides the stable-minimal-hypersurface regularity theorem and the $C^{1,\\alpha}$ compactness argument used in the appendix.","marker":"[32]"},{"why":"Provides the index-compactness lemma: disjoint negative-index regions force total index to grow, and index passes to limits.","marker":"[30]"},{"why":"Supplies the strong Frankel theorem for shrinkers: no proper $F$-stationary rectifiable hypersurface with $H^{n-2}(\\operatorname{sing} V)=0$ is $L$-stable.","marker":"[13]"},{"why":"Supplies the quantitative singular-set estimates that yield the Minkowski dimension $n-7$ bound in Theorem 6.8.","marker":"[3]"},{"why":"Supplies the quantitative stratification machinery and the definitions of singular strata used in the regularity-scale estimates.","marker":"[9]"},{"why":"Supplies the volume bounds for quantitative singular strata of stationary varifolds used in Lemma 6.6.","marker":"[29]"}],"fun_headline_variants":["In 3–6D, bounded curvature and index prevent MCF blow-up","Blow-up forces curvature or Morse index to diverge","Bounded mean curvature and index rule out singularities","No blow-up: bounded curvature and index keep MCF smooth","Singularity requires either curvature or index to explode"],"cache_read_input_tokens":27776,"weakest_assumption_plain":"The proof depends on Theorem 2.6, a two-sided pseudolocality estimate for flows with uniformly bounded mean curvature, which is stated and used without proof; if that estimate is false, curvature could build up away from the presumed singular set and the convergence to a stable limit cone would break down.","fun_headline_variants_meta":{"raw":{"variants":["In 3–6D, bounded curvature and index prevent MCF blow-up","Blow-up forces curvature or Morse index to diverge","Bounded mean curvature and index rule out singularities","No blow-up: bounded curvature and index keep MCF smooth","Singularity requires either curvature or index to explode"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000228,"raw_usage":{"total_tokens":1423,"prompt_tokens":842,"completion_tokens":581,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":458,"completion_tokens_details":{"reasoning_tokens":496}},"tokens_in":458,"tokens_out":581,"duration_ms":5998,"temperature":1.0,"reasoning_tokens":496,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T21:22:00.218134+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A concrete search: look for a smooth embedded mean curvature flow with uniformly bounded mean curvature whose time-zero slice has area ratio $1+o(1)$ in a ball, but whose second fundamental form exceeds $1/(\\varepsilon r_0)$ inside the parabolic cylinder predicted by Theorem 2.6. Finding one would disprove the unproved pseudolocality estimate and collapse the proof's foundation; proving Theorem 2.6 would close the paper's only explicit gap.","supporting_citations":[{"cited_title":"Huisken, Asymptotic behavior for singularities of t he mean curvature flow, J","cited_arxiv_id":null,"evidence_quote":"Provides the Gaussian monotonicity formula, density bounds, and the distance estimate used at the final rescaling to show regularity."},{"cited_title":"Li and B","cited_arxiv_id":null,"evidence_quote":"Supplies the pseudolocality lemmas, the decomposition of space into high-curvature, thick, and thin parts, and the overall multiplicity-one strategy."},{"cited_title":"Li and B","cited_arxiv_id":null,"evidence_quote":"Supplies the height-difference parabolic equation, the parabolic Harnack estimates, and the extraction of a positive solution on the limit cone."},{"cited_title":"Schoen and L","cited_arxiv_id":null,"evidence_quote":"Provides the stable-minimal-hypersurface regularity theorem and the $C^{1,\\alpha}$ compactness argument used in the appendix."},{"cited_title":"Sharp, Compactness of minimal hypersurfaces with bo unded index, J","cited_arxiv_id":null,"evidence_quote":"Provides the index-compactness lemma: disjoint negative-index regions force total index to grow, and index passes to limits."},{"cited_title":"A strong Frankel Theorem for shrinkers","cited_arxiv_id":"2306.08078","evidence_quote":"Supplies the strong Frankel theorem for shrinkers: no proper $F$-stationary rectifiable hypersurface with $H^{n-2}(\\operatorname{sing} V)=0$ is $L$-stable."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the quantitative singular-set estimates that yield the Minkowski dimension $n-7$ bound in Theorem 6.8."},{"cited_title":"Cheeger, R","cited_arxiv_id":null,"evidence_quote":"Supplies the quantitative stratification machinery and the definitions of singular strata used in the regularity-scale estimates."},{"cited_title":"Naber and D","cited_arxiv_id":null,"evidence_quote":"Supplies the volume bounds for quantitative singular strata of stationary varifolds used in Lemma 6.6."}],"review_version":1}