{"id":"c73c3e12-3b62-44d8-8ba1-7a7ad8e81e73","arxiv_id":"2512.19227","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For every k-dimensional Brakke flow there is a canonical Radon measure on spacetime times the Grassmannian whose disintegrations characterize all equivalent classical flows.","lead":"This mathematics paper constructs a canonical space-time-Grassmann measure for every Brakke flow, recording position, time, and tangent plane simultaneously, and proves this measure obeys the same evolution inequality as the original flow. Generalists should pay attention because it provides a new measure-based definition of Brakke flow that may unify convergence and compactness arguments in geometric measure theory.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem B's proof asserts without argument that inequality (29), shown only outside a null set, extends to all a,b∈J; the extension is needed for the claimed equivalence and is non-immediate because ⟨V,t⟩ is undefined on the null set.","rationale":"The abstract's central claim is the equivalence between classical Brakke flows and the new space-time measure definition. Theorem A (existence of λ) is a construction, but the equivalence itself rests on Theorem B, specifically on B.4 and B.5. The proof of B.4 contains an explicit unproved extension of inequality (29) from a full-measure set to all times. Because ∥⟨V,t⟩∥ is undefined on the null set N, this is not a routine continuity argument; it requires a careful limit using the left/right representatives established in B.2/B.3. The paper omits this argument, so the conclusion that ⟨V−,·⟩ and ⟨V+,·⟩ are Brakke flows is not fully justified as written. This is a concrete, localized gap in the proof of the paper's main equivalence. I considered the reader's other flagged concern, Lemma 3.6, which underpins Theorem A. The proof there is detailed and relies on Allard compactness and integration by parts; I could not identify a specific flaw. Therefore I do not treat Lemma 3.6 as the most load-bearing concern. The proposed concrete test is a finite derivation: write out the limit passage from (29) to the integrated Brakke inequalities. If successful, Theorem B is repaired; if not, the equivalence claim needs to be weakened or the proof expanded. My verdict remains CONDITIONAL, matching the reader's verdict, so no change is recommended.","tokens_in":15022,"tokens_out":15408,"duration_ms":138566,"concrete_test":"Write out the limit passage for (29). For a<b in J, choose sequences a_j∈(a,b)∩(J\\N) with a_j↓a and b_j∈(a,b)∩(J\\N) with b_j↑b. Use B.2/B.3 to identify lim_j ∥⟨V,a_j⟩∥(ϕ)=∥⟨V+,a⟩∥(ϕ) and lim_j ∥⟨V,b_j⟩∥(ϕ)=∥⟨V−,b⟩∥(ϕ), yielding ∥⟨V−,b⟩∥(ϕ)−∥⟨V+,a⟩∥(ϕ) ≤ ∫_a^b g. Then, using u_r≤u_l from the proof of B.2/B.3 to get ∥⟨V−,t⟩∥≥∥⟨V+,t⟩∥, deduce ∥⟨V−,b⟩∥(ϕ)−∥⟨V−,a⟩∥(ϕ) ≤ ∫_a^b g and ∥⟨V+,b⟩∥(ϕ)−∥⟨V+,a⟩∥(ϕ) ≤ ∫_a^b g. If this derivation succeeds, the gap is closed; if any step fails, Theorem B's conclusion that the one-sided representatives are Brakke flows is unsupported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"In the proof of Theorem B, after establishing B.2/B.3, the text states: 'Note that the inequality (29) can be extended for every a,b∈J, B.4 follows from B.2 and B.3.' Inequality (29) was proved only for a,b∈J\\N, where N is the null set on which the slicing limits ⟨V,t⟩, ⟨V−,t⟩, and ⟨V+,t⟩ fail to exist or disagree. For t∈N the expression ∥⟨V,t⟩∥(ϕ) in (29) is undefined, so the literal inequality cannot hold for every a,b. A correct extension would require passing through the left/right continuous representatives from B.2/B.3, producing mixed inequalities such as ∥⟨V−,b⟩∥(ϕ)−∥⟨V+,a⟩∥(ϕ) ≤ ∫_a^b g, and then deriving the integrated Brakke inequality for ∥⟨V−,·⟩∥ and ∥⟨V+,·⟩∥ separately. This argument is omitted. Because B.4 and B.5 are the core of the claimed equivalence between classical Brakke flows and the new space-time measure definition, the central claim is only conditionally supported unless this step can be filled. I also considered Lemma 3.6; its proof is detailed and I did not locate a concrete gap, so I do not press that concern here.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper defines, for a k-dimensional Brakke flow on an open set U over an open time interval J, a canonical Radon measure λ on J × G_k(U) by integrating the associated integral varifolds V(t) in time (Theorem A). It then studies arbitrary Radon measures V on J × G_k(U) satisfying a distributional Brakke inequality (1). Theorem B asserts that such measures correspond to equivalence classes of classical Brakke flows, with canonical left- and right-continuous representatives ⟨V−,t⟩ and ⟨V+,t⟩, and that any disintegration between these representatives is itself a Brakke flow. Theorem C asserts ∥λ∥-measurability of the tangent map, mean curvature vector, and density along the flow. The technical core includes a new proof of upper semicontinuity of the Brakke variation (3.6), avoiding an approximation step from Ilmanen that the paper shows can fail (3.8), and an L^1 measurability result for t ↦ V(t) (3.10.8).","tokens_in":15367,"tokens_out":14767,"duration_ms":147435,"significance":"If the main theorems are correct, this is a substantial contribution. The space-time-Grassmann measure gives a canonical, coordinate-free object attached to a Brakke flow, and Theorem B offers a genuinely new spacetime definition of Brakke flow that packages the known BV-type behavior of Brakke flows into left/right representatives. The measurability results in Theorem C are useful for further analytic and geometric arguments. The proof of 3.6 is itself a valuable technical improvement, as it identifies and repairs a gap in a standard approximation argument. The paper is generally careful, uses appropriate modern geometric measure theory tools, and avoids fitted parameters or circular reasoning. However, the proof of Theorem B contains a specific omitted argument in the passage from inequality (29) to B.4, and the statement of the main distributional inequality (1) needs a precise interpretation outside a null set. The central equivalence is therefore conditionally supported rather than fully established.","major_comments":[{"comment":"Inequality (29) is proved only for a,b ∈ J \\ N, where N is the null set on which the slice limits defining ⟨V,t⟩, ⟨V−,t⟩, and ⟨V+,t⟩ fail to exist or disagree. The sentence 'Note that the inequality (29) can be extended for every a,b∈J' is asserted without proof. Moreover, for t∈N the expression ∥⟨V,t⟩∥(φ) is not defined, so the literal inequality cannot hold for every a,b. This is load-bearing: B.4 and B.5 require an integrated Brakke inequality for the left and right representatives. To fill the gap one must pass through the left/right continuous representatives from B.2/B.3, proving for example ∥⟨V−,b⟩∥(φ)−∥⟨V+,a⟩∥(φ) ≤ ∫_a^b B(∥⟨V,t⟩∥,φ) dt for a<b, then deriving the corresponding inequalities for ∥⟨V−,·⟩∥ and ∥⟨V+,·⟩∥ separately using the ordering ∥⟨V−,t⟩∥ ≥ ∥⟨V+,t⟩∥. This argument is absent and is needed for the claimed equivalence between the new spacetime definition and classical","section":"§4, proof of Theorem B (Proof of B.4)"},{"comment":"The distributional inequality (1) contains B(∥⟨V,t⟩∥, φ), but ⟨V,t⟩ and hence ∥⟨V,t⟩∥ are defined only when the symmetric limit exists, which may fail on a null set N. As stated, (1) is not well-defined for an arbitrary Radon measure V. The intended meaning is presumably that the integrand is defined for L^1-a.e. t, with B allowed to be −∞ only on a null set, or that one should work with the left/right representatives. This precision is necessary because the proof of B.1 uses (1) to identify the distributional derivative of R with a Radon measure.","section":"Theorem B, Eq. (1)"},{"comment":"The step 'By 3.6, {t:V(t)(α)≥r} ∩ A(i) is relatively closed in A' is very terse. It requires using uniqueness of the integral varifold with a given weight to identify the limit varifold obtained from 3.6 with V(t) at continuity points of μ. A short explanation would substantially improve readability and make the L^1 measurability of V(t) — on which Theorem A depends — easier to verify. I do not see a fatal gap, but the argument should be spelled out.","section":"§3, proof of 3.10.8"}],"minor_comments":[{"comment":"The statement appears garbled: 'μ(t) = lim_{s↑t} μ(s)(ψ)' should presumably read 'μ(t)(ψ) = lim_{s↑t} μ(s)(ψ)'.","section":"§3, 3.10.5"},{"comment":"There is a duplicated heading 'Proof of B.1' immediately before 'Proof of B.1, B.2 and B.3'; the second heading should be 'Proof of B.2 and B.3'.","section":"§4, proof of B.1-B.3"},{"comment":"Minor typos: 'uniuq' in the abstract, 'characrterized' in Definition 2.4, 'Brakkfe' in Remark 3.12. The notation 'dmnV' is used without definition; it should be introduced.","section":"Throughout"},{"comment":"The one-sided limit statements involve the same letter t as the evaluation point, e.g. 'lim_{t↓a} ∥⟨V−,t⟩∥'. This is confusing; using a different dummy variable (e.g. s↓a) would clarify.","section":"Theorem B, B.2/B.3"},{"comment":"The final sentence 'the conclusion then follows by approximating I by its compact subsets and 4.1.4' likely should refer to 4.1.2 or 4.1.3; the reference is unclear.","section":"§4, proof of 4.1.4"}],"recommendation":"major_revision","confidential_remarks":"The B.4 gap is genuine but appears repairable by a standard limiting argument using the left/right representatives. The rest of the architecture is sound and the paper makes a valuable contribution. I recommend major revision rather than rejection: the authors should supply the missing extension of (29) and clarify the a.e. interpretation of (1)."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The contribution is real. For any Brakke flow the authors construct a canonical Radon measure λ on J×G_k(U) by integrating the canonical integral varifolds V(t) against Lebesgue measure, and they show the flow can be characterized by a distributional inequality. Theorem A is new, and the construction relies on a genuine improvement, Lemma 3.6, which fixes a flaw in Ilmanen's upper-semicontinuity argument. The measurability results in Theorem C are also new, and the proofs look sound. I read Lemma 3.6 carefully and did not find a gap.\n\nThe soft spot is in the proof of B.4. The paper states \"the inequality (29) can be extended for every a,b∈J\" but gives no argument. Inequality (29) is established only for a,b outside the null set N where the slicing limits exist. For t∈N the expression ∥⟨V,t⟩∥ in (29) is undefined, so the literal inequality cannot simply be extended. A correct extension would need to pass through the left- and right-continuous representatives from B.2/B.3, proving mixed inequalities such as ∥⟨V−,b⟩∥(ϕ)−∥⟨V+,a⟩∥(ϕ) ≤ ∫_a^b g and then deriving the integrated Brakke inequality for each representative separately. That argument is missing. Since B.4 and B.5 are the core of the claimed equivalence with classical Brakke flows, this is a load-bearing gap. It looks repairable—the ingredients are in the paper—but as written the central theorem is not fully proven.\n\nMinor observations: the paper is dense and relies on many external deep results; the authors are transparent about that. The new definition 4.11 is clean, and the left/right representative picture of jumps is appealing. The citation pattern is honest; related work by Hensel-Laux and Buet et al. is cited and distinguished.\n\nWho gets value: geometric measure theorists working on Brakke flow, compactness, and varifold solutions. The construction of λ and Lemma 3.6 are worth having even if Theorem B needs repair. I would send this to a serious referee. If the B.4 gap can be filled, this is a substantial contribution; if not, the equivalence theorem as stated doesn't follow.","headline":"Genuinely new canonical space-time-Grassmann measure for Brakke flow, but Theorem B has a load-bearing gap in the proof of B.4 that is likely fixable.","tokens_in":15832,"tokens_out":2507,"would_cite":true,"duration_ms":24358,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["53E10","28A75"],"pacs":[],"model":"deepseek-v4-flash","headline":"Every Brakke flow determines a canonical space-time-Grassmann measure λ, and this measure alone yields a new, equivalent definition of the flow.","keywords":["Brakke flow","mean curvature flow","space-time-Grassmann measure","varifolds","integral varifolds","upper semicontinuity","L1 measurability","BV functions"],"falsifier":"Exhibit a sequence of integral varifolds V_i with uniformly bounded mass on a set {φ>0}, converging to V, for which limsup_i B(∥V_i∥,φ) > B(∥V∥,φ); such an example would disprove Lemma 3.6 and with it the measurability of t↦V(t) and the existence of λ.","tokens_in":14909,"feed_emoji":"📐","tokens_out":6088,"duration_ms":54588,"temperature":0.7,"pith_summary":"Brakke flows are families of measures moving by mean curvature, traditionally tracked as a time-parametrized collection of varifolds. This paper proves that any such flow can be repackaged without loss into a single Radon measure λ on the space-time-Grassmannian, built by integrating the flow's varifolds against one-dimensional Lebesgue measure in time. It then shows that a measure satisfying a simple distributional inequality is exactly the same object as a classical Brakke flow, with left- and right-continuous representatives that pin down the flow's jump discontinuities. Finally, it proves that the tangent plane, mean curvature vector, and density are all measurable with respect to λ's weight measure, so geometric data can be integrated directly over space-time. If sound, this gives Brakke-flow theory a canonical measure-theoretic foundation paralleling the one varifolds give to stationary surfaces.","feed_headline":"One measure now defines every Brakke flow","feed_subtitle":"A Radon measure over space and k-planes encodes the whole flow, matching the classical definition exactly.","key_machinery":"The central object is the space-time-Grassmann measure λ, defined by integrating the flow's integral varifolds V(t) against time-Lebesgue measure. Its construction hinges on a new proof (3.6) of the upper semicontinuity of the Brakke variation B(∥V∥, φ) under varifold convergence, which avoids an interior approximation shown to fail in general (3.8) and instead uses the divergence theorem and second-order rectifiability. Theorem B then uses BV-function theory to extract left/right representatives from the space-time measure.","core_discovery":"We show that for any k-dimensional Brakke flow {μ(t)} on U over J, the map t↦V(t) sending each time to the unique integral varifold with weight μ(t) is L^1-measurable, so we can define a Radon measure λ on J×G_k(U) by λ(f)=∫∫ f(t,x,S) dV(t)(x,S) dL^1_t. We then prove that a Radon measure V over J×G_k(U) satisfying the distributional inequality (1) is equivalent to a classical Brakke flow: the left and right derivatives of its disintegrations define left- and right-continuous Brakke flows that sandwich every equivalent disintegration, and these representatives are themselves Brakke flows. Consequently the space-time measure formulation can serve as a new definition of Brakke flow, equivalent","pith_inferences":["This suggests a parabolic counterpart to the varifold compactness theorem: a sequence of space-time measures satisfying (1) may converge to a measure of the same type, giving a compactness framework for Brakke flows directly in space-time.","The left/right representative construction may yield a canonical way to continue a Brakke flow through a jump, choosing the right-continuous representative as the physically forward evolution.","One could test the framework on explicit solutions, such as a shrinking sphere, to compute λ explicitly and verify the measurability and inequality directly.","The Borel measurability of the tangent map on varifold space suggests that the space-time measure λ could support a tangent-measure decomposition at (t,x) analogous to Allard's tangent varifolds, perhaps yielding partial regularity."],"forward_implications":["Every Brakke flow now has a canonical space-time-Grassmann measure, allowing the flow to be treated as a single measure rather than a time-parametrized family.","The distributional inequality (1) gives a new definition of Brakke flow that is provably equivalent to the classical ones, so any theorem about classical Brakke flows transfers to this formulation.","The left- and right-continuous representatives isolate and describe jump discontinuities of the flow, which occur exactly where the representatives disagree.","The measurability of tangent map, mean curvature, and density with respect to ∥λ∥ permits integrals of these geometric quantities over space-time, not just slice by slice."],"fun_headline_variants":["One measure now defines every Brakke flow","Space-time measure encodes all Brakke flows","A single Radon measure captures any Brakke flow","New measure unifies Brakke flow definitions","Brakke flow redefined by one space-time measure"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The existence of λ depends on the L^1 measurability of t↦V(t), which rests on the new upper-semicontinuity lemma (3.6); if that lemma fails, the whole construction collapses.","fun_headline_variants_meta":{"raw":{"variants":["One measure now defines every Brakke flow","Space-time measure encodes all Brakke flows","A single Radon measure captures any Brakke flow","New measure unifies Brakke flow definitions","Brakke flow redefined by one space-time measure"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000146,"raw_usage":{"total_tokens":1007,"prompt_tokens":720,"completion_tokens":287,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":464,"completion_tokens_details":{"reasoning_tokens":212}},"tokens_in":464,"tokens_out":287,"duration_ms":3503,"temperature":1.0,"reasoning_tokens":212,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-03T14:45:02.495341+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Exhibit a sequence of integral varifolds V_i with uniformly bounded mass on a set {φ>0}, converging to V, for which limsup_i B(∥V_i∥,φ) > B(∥V∥,φ); such an example would disprove Lemma 3.6 and with it the measurability of t↦V(t) and the existence of λ.","supporting_citations":[],"review_version":1}