{"id":"28b3e70b-da37-42a6-8818-8d0528dba8f8","arxiv_id":"1908.03033","paper_version":3,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"If both sides of an Ahlfors regular boundary have Poisson kernels with logarithms in VMO, then the domain is a vanishing chord-arc domain, and conversely.","lead":"This paper proves that a domain with a uniformly thick boundary is flat in the limit exactly when the logarithms of the interior and exterior Poisson kernels have vanishing mean oscillation. This gives a clean potential-theoretic fingerprint of smooth boundary with minimal assumptions, only connectedness and Ahlfors regularity.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Lemma 4.3 is the sole bridge from local doubling of harmonic measure to interior corkscrews, and the paper only sketches it; a circular adaptation of [HM15, Lemma 3.14] would leave Theorem 4.12, and hence Theorem 1.1 (ii)->(i), unproved.","rationale":"The reader's weakest_assumption correctly identifies Lemma 4.3. My independent pass agrees: the other parts of the proof are detailed (Section 3 and Appendix A are written out; Theorem 4.12 follows the [BH16] template with constants tracked), so the point where the analytic-to-geometric step is least secure is precisely the sketched folk lemma. I am not claiming Lemma 4.3 is false; the sketch is plausible and the authors credit S. Hofmann for the argument. But the paper explicitly says the proof is a sketch with a slight modification of [HM15, Lemma 3.14], and it is the unique step that produces the two-sided corkscrew used to build the UR approximations T_Q^pm in Appendix B. If the adaptation in (4.3)-(4.6) is circular or requires an unstated geometric hypothesis, the implication (ii)->(i) is incomplete. That is a verification gap at a load-bearing point, not a demonstrated contradiction, so I move the verdict from unconditional acceptance to conditional acceptance pending the concrete check.","tokens_in":43670,"tokens_out":41647,"duration_ms":439005,"concrete_test":"Write out the full proof of Lemma 4.3 without citing [HM15, Lemma 3.14], and verify the step from (4.5) to (4.6) line by line: for each Whitney cube I in the boundary strip with center Y_I and an associated boundary point x_I, derive G(Y_I) <= C (ell(I)/r)^alpha (1/|2B|) int_{2B cap Omega} G using only Ahlfors regularity, boundary Holder continuity of the Green function, and local doubling of harmonic measure, then check that the dyadic summation constant is uniform in rho and in the compact set. Record every use of Harnack or boundary Harnack; if the derivation requires an interior corkscrew at scale r, or any geometric conclusion that Lemma 4.3 is meant to establish, the argument is circular and Theorem 4.12 lacks its corkscrew input. If the derivation is non-circular and the constants are uniform, the concern does not land.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"Theorem 1.1 (ii)->(i) runs through Theorem 4.12, whose proof uses Lemma 4.3 to convert local doubling of omega_+ and omega_- into interior corkscrew points, and then uses those corkscrews as the DLTSCS input for the UR approximating domains T_Q^pm in Appendix B. Lemma 4.3 is therefore load-bearing: without it the L^p estimates and jump relations for single layer potentials in (4.12), (4.38), and (4.39) have no domain to attach to. The paper itself flags the fragile spot: the lemma is called folklore, its proof is described as a sketch, and it is said to require 'a slight modification of the argument in [HM15, Lemma 3.14]'. The critical passage (4.5)-(4.6) compares G(Y_I) with r omega(CB)/sigma(CB) for Whitney cubes I in the boundary strip, invoking boundary Holder continuity and (4.3); the supporting geometric summation is postponed and only sketched. What must be checked is whether that comparison, and the Whitney summation, use only Ahlfors regularity, boundary Holder continuity of the Green function, and local doubling, or whether a hidden boundary Harnack/corkscrew input at scale r is assumed. Such a hidden geometric input would make the argument circular and break the chain from log k in VMO to nu in VMO.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proves Theorem 1.1: for n ≥ 3 and complementary domains Ω+ and Ω− = Rn \\ Ω+ with common (n−1)-Ahlfors regular boundary ∂Ω, the two domains are both vanishing chord-arc domains with ν ∈ VMOloc(σ) if and only if the interior and exterior Poisson kernels exist and log k+ and log k− lie in VMOloc(dσ). The forward direction is imported from Kenig–Toro [KT03], while the reverse direction is obtained by combining a new geometric flatness result (Corollary 3.10), a localization theorem for Poisson kernels (Theorem 4.12), and an approximation of domains satisfying a local two-sided corkscrew condition by UR domains (Appendix B). Section 3 develops excess and flatness methods for sets of locally finite perimeter with Ahlfors regular boundary, and the paper also states quantitative versions in Theorems 4.12 and 4.14.","tokens_in":43987,"tokens_out":4154,"duration_ms":49061,"significance":"If correct, Theorem 1.1 gives a sharp potential-theoretic characterization of vanishing chord-arc geometry under minimal topological assumptions, removing a priori NTA, chord-arc, or uniform rectifiability hypotheses that appear in [KT06] and [BH16]. The geometric results in Section 3 and in Appendices A and B are presented with substantial detail, and Theorem 4.12 provides explicit quantitative control of the oscillation of the unit normal in terms of the oscillation of log k±. The paper therefore represents a significant advance in the two-phase free boundary problem for harmonic measure. However, the reverse implication relies critically on Lemma 4.3, whose proof is only a sketch, and on an imported estimate in (4.22), and those points leave the proof not fully self-contained at a load-bearing location.","major_comments":[{"comment":"Lemma 4.3 is load-bearing for Theorem 4.12 and hence for the (ii) implies (i) direction of Theorem 1.1, but its proof is only sketched. The critical estimate (4.6) is obtained by an 'elementary geometric argument' that is postponed and then only outlined, and the lemma is called folklore. As written, the argument uses boundary Hölder continuity of the Green function and the estimate (4.3), but it does not fully demonstrate that the Whitney summation in (4.5)-(4.6) uses only Ahlfors regularity, boundary Hölder continuity, and local doubling of ω. If the comparison at (4.5)-(4.6) secretly requires a boundary Harnack estimate or a corkscrew point at scale r, the argument would be circular, because producing those corkscrews is precisely the role of the lemma. Please provide a complete proof, or a precise reference with the full statement, and verify explicitly that every input is available before the corkscrew conclusion is drawn.","section":"Section 4.1, Lemma 4.3"},{"comment":"The estimate (4.22), which is used in (4.38) to control the term I in Theorem 4.12, is quoted from [BH16, Lemma 1.33] without proof. This inequality converts the BMO bound on log k into the L2 closeness of k to its geometric mean, and it is a load-bearing step in the localization argument from log k± ∈ VMOloc(σ) to ν ∈ VMOloc(σ). Please include the proof or a full statement with hypotheses, constants, and the precise way in which the bound depends on the BMO seminorm, rather than a parenthetical reference, so that the reader can verify that the chain is complete.","section":"Section 4.2, equation (4.22)"},{"comment":"The proof of Theorem 1.1 concludes that (ii) implies (i) by combining Theorem 4.12 with Corollary 3.11. Since Theorem 4.12 depends on Lemma 4.3 and on (4.22), the main theorem inherits any incompleteness in those two points. In addition, Theorem 4.14 is stated as a direct consequence of Theorem 4.12, Corollary 3.10, and [KT99], but the dependence of the constants on the compact set is not made fully explicit; please clarify that the quantitative statement is uniform on compacta in the sense of the definitions of BMOloc and VMOloc.","section":"Section 4.3, Theorem 1.1 and Theorem 4.14"}],"minor_comments":[{"comment":"The manuscript contains numerous OCR-style artifacts, including 'iintegdisplay' in (4.2), 'nelementF' in Remark 4.6, and a corrupted running header 'FLA TNESS AND OSCILLA TION'. These should be cleaned before publication.","section":"Throughout"},{"comment":"In the proof of the claim in Remark 2.21, the notation mixes ∂∗E and ∂E several times, for example in (2.23) and in the line following (2.18). Please standardize the notation so that the role of the reduced boundary versus the topological boundary is unambiguous.","section":"Remark 2.21"},{"comment":"In part (4) of Lemma 4.11 the parameter is denoted '~τ(r)', which is visually confusing given the use of τ elsewhere. A more standard notation such as τ_r or τ(r) would improve readability.","section":"Lemma 4.11"},{"comment":"The statement 'there exists a dyadic cube Q as in Lemma B.2 such that Δ(x0, r0/A) ⊂ Q ⊂ Δ(x0, r0)' should define the surface ball Δ(x0, r) and the precise sense in which the cube Q contains one surface ball and is contained in another; this is currently implicit and would be easier to check if written out.","section":"Theorem 4.12, beginning of proof"}],"recommendation":"major_revision","confidential_remarks":"The main theorem is likely correct and the geometric material in Sections 3 and Appendices A-B is carefully written. My recommendation is driven by the fact that Lemma 4.3, which is the sole bridge from doubling of harmonic measure to interior corkscrews, is only sketched, and by the fact that (4.22) is imported without proof. These are fixable within the scope of the manuscript, but they are load-bearing for the central claim, so I do not feel comfortable accepting the paper in its current form."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is the theorem the field wanted, and it is mostly proved. The new content is real—Section 3 extends De Giorgi's flatness-via-excess technology from perimeter minimizers to arbitrary sets of locally finite perimeter with Ahlfors regular boundary, and that extension is what lets the paper replace chord-arc or uniform rectifiability hypotheses by mere connectedness. Corollary 3.10 and Theorem 4.12 are the core, and Appendix B's UR approximation is a useful piece of machinery on its own. The equivalence in Theorem 1.1 is not in Kenig-Toro or Bortz-Hofmann, and the authors are honest about what they import.\n\nWhere I would push: Lemma 4.3. It is the bridge from local doubling of harmonic measure to interior corkscrew balls, and it is presented as a folklore result whose proof is a small modification of [HM15, Lemma 3.14], with the Whitney summation sketched and postponed. Both the L^p layer-potential estimates and the jump relations in Theorem 4.12 sit on top of those corkscrews via the DLTSCS condition in Appendix B, so a hidden geometric assumption in the Green function comparison (4.3)–(4.6) would be circular. I did not see an explicit hidden assumption; the cited bound is claimed to depend only on n and the AR constant, and if that is correct the argument is a straightforward absorption. But the sketch is load-bearing enough that a referee should verify the adaptation in detail rather than take it on faith.\n\nMinor: (i)->(ii) is imported from [KT03], and (4.22) from [BH16]; both are clearly cited. The self-citations are background, not a circular load. No invented entities, no free parameters.\n\nWho this is for: anyone working on harmonic measure, two-phase free boundaries, or quantitative GMT. Section 3 alone is worth a read. This should go to peer review. I would not desk reject it; I would send it out and ask specifically for a line-by-line check of Lemma 4.3 and the HM15 comparison.","headline":"A real two-phase characterization theorem: Poisson kernel VMO is equivalent to vanishing chord-arc under just Ahlfors regularity and connectedness, with the main load-bearing soft spot being the sketched corkscrew lemma.","tokens_in":44504,"tokens_out":5035,"would_cite":true,"duration_ms":65123,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35R35","49J52","28A75","31A15"],"pacs":[],"model":"deepseek-v4-flash","headline":"Vanishing oscillation of the two Poisson kernel logarithms fully characterizes vanishing chord-arc geometry on Ahlfors regular boundaries.","keywords":["two-phase free boundary problem","harmonic measure","Poisson kernel","vanishing chord-arc domain","VMO","Ahlfors regular boundary","Reifenberg flatness","uniform rectifiability"],"falsifier":"Construct a connected domain with Ahlfors regular boundary and locally doubling harmonic measure whose boundary has no interior corkscrew ball at some arbitrarily small scale, for instance a boundary with hairpin crevices of decreasing width; if such a domain exists, Lemma 4.3 is false and the derivation of corkscrews from doubling, and hence Theorem 1.1, collapses.","tokens_in":43492,"feed_emoji":"📐","tokens_out":7207,"duration_ms":75733,"temperature":0.7,"pith_summary":"This paper tries to prove that the geometry of a rough two-sided domain boundary is completely encoded in the harmonic-measure data on the two sides. More precisely, for a connected domain and its connected complement in R^n with n at least 3, sharing an Ahlfors regular boundary, the two domains are vanishing chord-arc domains with unit normal in VMO if and only if the interior and exterior Poisson kernels exist and their logarithms have vanishing mean oscillation. The point is that no a priori flatness, chord-arc, or topological regularity is assumed; the potential-theoretic condition alone forces the boundary to be uniformly rectifiable, Reifenberg flat at vanishing scales, and to have two-sided corkscrew access. A sympathetic reader cares because this is a two-phase free boundary statement: a purely analytic condition on both Poisson kernels, which is scale-invariant and measurable in practice, fully determines the fine geometry of the interface.","feed_headline":"Two Poisson kernel logs characterize vanishing chord-arc domains","feed_subtitle":"For connected domains with Ahlfors regular boundary, the analytic condition alone forces full vanishing chord-arc geometry.","key_machinery":"The argument runs through three linked mechanisms. First, the De Giorgi cylindrical excess e(E,x,r,nu) = (1/r^(n-1)) integral over the cylinder of |nu_E - nu|^2/2 with respect to surface measure measures how far the boundary's measure-theoretic normal deviates from a fixed direction; a compactness-and-height-bound argument (Theorem 3.9 and Appendix A) shows that small excess forces the boundary to be a Lipschitz graph over a plane with controlled height, so control of the unit normal oscillation yields Reifenberg flatness. Second, on the potential-theoretic side, the small-BMO condition on log k+ and log k- implies, through the John-Nirenberg inequality, that k+ and k- are Muckenhoupt weights, which gives the reverse Holder and doubling properties of harmonic measure (Lemma 4.11); doubling then produces interior and exterior corkscrew balls (Lemma 4.3), hence uniform rectifiability. Third, using approximate domains built from dyadic cubes (Appendix B), the proof compares the single layer potentials on the two sides, applies the jump relations for their gradients, and transfers oscillation of the Poisson kernels to oscillation of the unit normal (Theorem 4.12) via Calderon-Zygmund estimates on the uniformly rectifiable approximations.","core_discovery":"The central discovery is Theorem 1.1: for n at least 3, if $\\Omega$+ and $\\Omega$- = R^n \\ $\\Omega$+ are domains with common topological boundary that is (n-1)-Ahlfors regular, then (i) both $\\Omega$+ and $\\Omega$- are vanishing chord-arc domains with unit normal in VMO_loc if and only if (ii) there exist poles X+ in $\\Omega$+ and X- in $\\Omega$- such that the Poisson kernels k+ = domega+/dsigma and k- = domega-/dsigma exist and log k+ and log k- lie in VMO_loc. The forward direction is the existing characterization of chord-arc domains by Poisson kernels; the new content is the reverse direction, which manufactures all the geometric information two-sided corkscrew balls, uniform rectifiability, vanishing Reifenberg flatness, and VMO of the unit normal out of the vanishing oscillation of the two logarithms, under only connectivity and Ahlfors regularity.","pith_inferences":["I infer that the quantitative estimates of Theorem 4.12 should transfer to a purely local statement: on any boundary ball, a sufficiently small BMO norm of log k+ and log k- controls the BMO norm of nu on a smaller ball, with the ratio of scales depending on the compact set, so the equivalence may hold for boundaries that are only Ahlfors regular up to a finite scale.","I infer that the same strategy may prove an analogous two-phase characterization for elliptic measures of uniformly elliptic operators with rough coefficients, since the corkscrew-from-doubling step is purely potential-theoretic and does not use the specific form of the Laplacian beyond the Green-function estimates.","I infer that the exponent one-eighth appearing in the oscillation transfer is likely an artifact of the John-Nirenberg and Calderon-Zygmund chain, and an explicit family of examples could reveal the true optimal exponent relating BMO norms of log k and nu."],"forward_implications":["If log k+ and log k- lie in VMO_loc, then the unit normal nu lies in VMO_loc and the common boundary is vanishing Reifenberg flat, so both domains are vanishing chord-arc domains.","Under the same hypotheses the boundary is uniformly rectifiable, so L^2 Riesz transforms are bounded and singular-integral tools apply to the two-phase problem.","The quantitative version shows that a sufficiently small BMO norm of log k+ and log k- forces the domains to be delta-chord-arc for arbitrarily small delta, with the required smallness depending only on dimension, the Ahlfors constant, and the pole locations.","Conversely, if the domains are already vanishing chord-arc, the logarithms of the Poisson kernels are in VMO_loc, so the potential-theoretic condition is not merely sufficient but exactly equivalent to the geometric one.","Because connectivity and Ahlfors regularity are the only a priori hypotheses, the result removes the need to assume Reifenberg flatness or a two-sided John condition in related free-boundary theorems."],"supporting_citations":[{"why":"This lemma supplies the argument that local doubling of harmonic measure yields interior corkscrew balls, which the paper adapts in Lemma 4.3.","marker":"[HM15, Lemma 3.14]"},{"why":"This work provides the singular-integral scheme through which small BMO of the Poisson kernel logarithms is converted into oscillation of the unit normal.","marker":"[BH16]"},{"why":"This paper establishes the forward direction, that vanishing chord-arc domains have Poisson kernel logarithms in VMO_loc.","marker":"[KT03]"},{"why":"This work supplies the quantitative converse and free-boundary regularity results used in Theorem 4.14.","marker":"[KT99]"},{"why":"This reference provides the nontangential maximal estimates and jump relations for single layer potentials on uniformly rectifiable domains used in Lemma 4.8 and Lemma 4.9.","marker":"[HMT10]"},{"why":"This book supplies the De Giorgi excess, compactness, and height-bound toolkit used throughout Section 3 and Appendix A.","marker":"[Mag12]"}],"fun_headline_variants":["Vanishing chord-arc from VMO logs of Poisson kernels","Converse: VMO Poisson logs force vanishing chord-arc geometry","Poisson kernel logs with VMO yield vanishing chord-arc domains","From VMO logs of Poisson kernels to vanishing chord-arc geometry"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The chain depends on the step in Section 4.1, Lemma 4.3, sketched rather than fully proved, that a locally doubling harmonic measure on a domain with Ahlfors regular boundary forces, at every boundary point and small scale, a ball inside the domain of size comparable to the scale; if that step fails, the analytic VMO condition cannot be turned into the two-sided boundary access the rest of the proof needs.","fun_headline_variants_meta":{"raw":{"variants":["Vanishing chord-arc from VMO logs of Poisson kernels","Converse: VMO Poisson logs force vanishing chord-arc geometry","Poisson kernel logs with VMO yield vanishing chord-arc domains","From VMO logs of Poisson kernels to vanishing chord-arc geometry"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000595,"raw_usage":{"total_tokens":2728,"prompt_tokens":830,"completion_tokens":1898,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":446,"completion_tokens_details":{"reasoning_tokens":1824}},"tokens_in":446,"tokens_out":1898,"duration_ms":16045,"temperature":1.0,"reasoning_tokens":1824,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T14:27:29.629088+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Construct a connected domain with Ahlfors regular boundary and locally doubling harmonic measure whose boundary has no interior corkscrew ball at some arbitrarily small scale, for instance a boundary with hairpin crevices of decreasing width; if such a domain exists, Lemma 4.3 is false and the derivation of corkscrews from doubling, and hence Theorem 1.1, collapses.","supporting_citations":[],"review_version":1}