{"id":"62c04c3f-9dfe-4289-8cd0-63d707c715a8","arxiv_id":"1908.09175","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Finite convex-hull width does not imply a Jordan curve is a quasicircle, but boundary width below cosh^{-1}(sqrt(2)) does.","lead":"The paper introduces a width invariant for Jordan curves in the Riemann sphere, defined through convex hulls in hyperbolic 3-space. It shows that width below an explicit constant forces a curve to be a quasicircle, while finite width does not.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Lemma 3.3, a key step in the proof of Theorem B, is false as stated: three vertical planes in the upper half-space model give a counterexample with distances below w0 but no intersection point.","rationale":"The reader's concern focused on the extremal reduction in Lemma 3.2. My independent check of the subsequent lemmas found a sharper problem: Lemma 3.3, which is used directly in Corollary 3.4 to prove the quasi-isometry of nearest-point projections, is false as stated. The counterexample is simple and uses only vertical geodesic planes in the upper half-space model. It satisfies exactly the hypotheses written in Lemma 3.3 (P and Q bound disjoint halfspaces; P′ and Q bound disjoint halfspaces), gives distances below w0, and has P∩P′=∅. Thus the claimed conclusion fails. The proof's appeal to Lemma 3.2 does not rescue it, since Lemma 3.2 assumes a stronger pairwise-disjoint-halfspaces condition, and even under that stronger condition two asymptotic planes need not intersect. Since the small-width theorem rests on this lemma, the central proof is invalid as written. The theorem might be salvageable with additional hypotheses (e.g., support planes of a common convex set), but the manuscript does not state or prove them. This moves the verdict from CONDITIONAL to REJECT for the current version.","tokens_in":13895,"tokens_out":21555,"duration_ms":232810,"concrete_test":"Verify the counterexample: fix upper half-space H3, Q={x=0}, P={x=-1}, P′={x=-2}; orient halfspaces H_Q={x>0}, H_P={x<-1}, H_P′={x<-2}. Set y=(0,0,10), x=(-1,0,10), x′=(-2,0,10). The Euclidean segments have hyperbolic length 1/10 and 2/10, so both hyperbolic distances are <0.881. P∩P′ is empty. This refutes Lemma 3.3. Then check whether Cor. 3.4's support planes of a convex hull can realize two parallel asymptotic supporting planes on the same side; if yes, Theorem 3.1 lacks proof; if no, the lemma needs an extra hypothesis and a new proof.","verdict_should_be":"REJECT","load_bearing_attack":"The proof of Theorem 3.1 (and hence Theorem B) relies on Lemma 3.3 to show the nearest-point projections π± are quasi-inverse quasi-isometries (Cor. 3.4, Lemma 3.6). But Lemma 3.3 is false as stated. In the upper half-space model, take Q={x=0}, P={x=-1}, P′={x=-2}, with halfspaces H_Q={x>0}, H_P={x<-1}, H_P′={x<-2}. Then P and Q bound disjoint halfspaces, and P′ and Q do too. For y=(0,0,Z), x=(-1,0,Z), x′=(-2,0,Z), the hyperbolic distances satisfy d(x,y)≤1/Z and d(x′,y)≤2/Z; for Z=10 these are well below w0=cosh^{-1}(√2). Yet P∩P′=∅, so no z∈P∩P′ exists. The proof's claim 'P∩P′ is non-empty by Lemma 3.2' is invalid: Lemma 3.2 requires all three planes to bound pairwise disjoint halfspaces, a hypothesis not present in Lemma 3.3; moreover even if added, asymptotic planes would still give no intersection. Therefore the small-width argument has a concrete gap.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper defines a hyperbolic width invariant w(C) and a boundary variant w_∂(C) for Jordan curves in CP^1, in analogy with the AdS width of Bonsante-Schlenker. It proves three main results: Theorem A gives a curve of finite width that is not a quasicircle; Theorem B states that every Jordan curve with w_∂(C) < cosh^{-1}(√2) is a quasicircle; and Theorem C characterizes quasicircles by the property that a nearest-point projection between the two boundary components of the convex hull is a quasi-isometry. Section 2 contains the construction for Theorem A, Section 3 proves Theorem B via a quantitative version (Theorem 3.1), Section 3.3 discusses optimality of the threshold, and Section 4 shows that w_∂ and w can differ.","tokens_in":14150,"tokens_out":12646,"duration_ms":129182,"significance":"The proposed invariant is natural and the contrast with the AdS characterization (Proposition 1.3) is conceptually interesting. The constructions in Section 2 are explicit and use Ahlfors' bounded turning criterion and a supporting-plane compactness argument, which are strengths. If the proof of Theorem B can be repaired, the explicit threshold w0 = cosh^{-1}(√2) and the optimality interval would be a clean quantitative statement. However, the central argument currently rests on a false lemma, so the significance is conditional on a substantial repair.","major_comments":[{"comment":"Lemma 3.3 is false as stated. In the upper half-space model, let Q={x=0} with half-space H_Q={x>0}, P={x=-1} with half-space H_P={x<-1}, and P'={x=-2} with half-space H_{P'}={x<-2}. Then P (resp. P') and Q bound disjoint half-spaces, but P∩P'=∅. Take y=(0,0,Z), x=(-1,0,Z), x'=(-2,0,Z). For Z=10 the hyperbolic distances satisfy d(x,y)<w0 and d(x',y)<w0, yet no z∈P∩P' exists. The proof's appeal to Lemma 3.2 is invalid because Lemma 3.2 assumes all three planes bound pairwise disjoint half-spaces, a hypothesis absent from Lemma 3.3. Since Corollary 3.4, Lemma 3.6, and hence Theorem 3.1 and Theorem B all rely on Lemma 3.3, this is a load-bearing gap.","section":"§3.1, Lemma 3.3"},{"comment":"The proof of Lemma 3.2 asserts that the worst case is obtained when the three lines are pairwise asymptotic and the point y is symmetrically placed, and that this reduction follows by symmetry. This reduction is not proved. Since the value w0 = cosh^{-1}(√2) is the basis for the threshold in Theorem B, a rigorous extremal argument or a reference is needed. As written, the calculation of the four congruent triangles is conditional on an unproved reduction, so Lemma 3.2 is not fully established.","section":"§3.1, Lemma 3.2"},{"comment":"In the converse direction of Proposition 3.7, the proof asserts that π+ extends continuously to the identity on C because 'π+ moves points at most by η0'. However, η0 comes from Lemma 3.8, which is proved only under the forward hypothesis that C is a quasicircle. Under the converse hypothesis, the bounded-distance estimate d(y,π+(y))≤η0 has not been established, and a quasi-isometry between ∂-C and ∂+C does not by itself guarantee that its boundary extension is the identity. The composition b+∘π+∘(b-)^(-1) therefore cannot yet be asserted to extend to the identity on C. This gap affects the conclusion that C is the image of RP^1 under a quasiconformal map and needs to be repaired, for example by proving from the quasi-isometry condition that π+ is asymptotically the identity near C.","section":"§3.2, Proposition 3.7"}],"minor_comments":[{"comment":"The notation is confusing because w0 was defined as cosh^{-1}(√2) at the start of Section 3, while the example in Proposition 3.9 concerns sinh^{-1}(√2). Please state explicitly that the example gives an upper bound on the optimal threshold, while Theorem B provides a lower bound.","section":"§3.3, Proposition 3.9"},{"comment":"In the proof of Proposition 2.4, the planes P^+ and P^- are introduced with boundaries in U^+ and U^-, but the sign convention is not defined. The inequality d(x_n,∂+CH(C_n)) = d(x̄,g_n(∂+CH(C_n))) < d(x̄,P^+) should be expanded, since it implicitly uses the fact that projection onto a convex set is distance-decreasing and that the chosen support planes separate the relevant regions.","section":"§2.2, Proposition 2.4"},{"comment":"Claim 2.6 is presented as evident from Figure 2. A short proof or a more precise statement of which complementary regions are involved would improve the exposition and make the verification of the hypotheses of Proposition 2.4 easier to check.","section":"§2.2.1, Claim 2.6"},{"comment":"In the proof of Lemma 3.8, the phrase 'the path metric on f(Σ_r) is L-bilipschitz to the path metric on Σ_r' should be stated more carefully as 'L-bilipschitz to the induced path metric on Σ_r', and the subsequent 'cosh(r)-bilipschitz embedded copy of H^2' should clarify the comparison between the induced path metric on Σ_r and the hyperbolic metric on H^2. There are also minor typos such as 'equiped' in the proof of Proposition 3.7.","section":"§3.2, Lemma 3.8"}],"recommendation":"major_revision","confidential_remarks":"The paper is within the scope of the journal and the topic is timely. The false Lemma 3.3 is a serious but possibly repairable gap in the proof of Theorem B. I recommend giving the authors an opportunity for major revision, asking them to repair Lemma 3.3 under the hypotheses actually available from the convex hull geometry, to justify the extremal reduction in Lemma 3.2, and to fix the boundary-extension issue in Proposition 3.7. If these points are resolved, the paper would make a solid contribution."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things to know. The paper introduces a useful invariant, the width of a Jordan curve in CP1, and proves a genuine counterexample (Theorem A): finite width does not imply quasicircle. That part is solid and novel. The second thing is that the proof of the positive theorem (Theorem B) has a real hole: Lemma 3.3 is false as stated.\n\nHere is the counterexample to Lemma 3.3. In the upper half-space model let Q={x=0}, P={x=-1}, P'={x=-2}, with halfspaces x>0, x<-1, x<-2. For y=(0,0,Z), x=(-1,0,Z), x'=(-2,0,Z), the hyperbolic distances are about 1/Z and 2/Z, both below w0=cosh^{-1}(sqrt(2)) for large Z. P and P' do not meet, so no z exists. The proof claims P∩P' is non-empty by Lemma 3.2, but Lemma 3.2 assumes three pairwise disjoint halfspaces, which is not a hypothesis of Lemma 3.3, and the example doesn't satisfy it. This is not a nitpick: Corollary 3.4, Lemma 3.6, and Theorem B all lean on Lemma 3.3. The support-plane setting in Corollary 3.4 might supply extra structure, but the paper doesn't say what it is.\n\nWhat the paper does well: Theorem A's construction is the real contribution. The width invariant is natural, the AdS analogy is clearly explained, and Theorem C—quasicircle iff nearest-point projection is a quasi-isometry—is a plausible and useful characterization argued with standard tools. The citation pattern is fine; self-citations are motivational, not load-bearing. I also found the Kerckhoff-Thurston example in Section 4 convincing.\n\nSofter spots: Lemma 3.2's reduction to the symmetric ideal triangle is terse; I would ask for a real proof. Proposition 3.9's symmetry calculation is also condensed. Both are minor compared to Lemma 3.3.\n\nBottom line: the paper deserves a referee, but the referee should be told to focus on Section 3. As is, Theorem B is unproven. If the authors can patch Lemma 3.3 or replace it with a correct statement that works in the support-plane context, the paper is solid. I would not desk-reject.","headline":"New width invariant and a solid counterexample in Theorem A, but the proof of Theorem B leans on a false Lemma 3.3; worth refereeing with Section 3 flagged.","tokens_in":14666,"tokens_out":10622,"would_cite":true,"duration_ms":110094,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["30C62","30F40","51M10","57M50"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper proves that finite convex-hull width does not force a Jordan curve to be a quasicircle, but sufficiently small width does, with an explicit threshold.","keywords":["quasicircles","Jordan curves","convex hull","hyperbolic 3-space","width of curves","quasiconformal maps","nearest point projection","anti-de Sitter geometry"],"falsifier":"Exhibit a Jordan curve that is not a quasicircle and whose boundary width is strictly less than $\\cosh^{-1}(\\sqrt{2})$; Theorem B forbids it. A more targeted check: numerically optimize over triples of disjoint half-spaces in $\\mathbb{H}^3$ the quantity $\\max(d(x,y),d(x',y))$ as in Lemma 3.2; a value below $w_0$ would break the proof of Theorem B.","tokens_in":13711,"feed_emoji":"📐","tokens_out":10164,"duration_ms":79648,"temperature":0.7,"pith_summary":"This paper studies a geometric invariant called width, defined for any Jordan curve in the Riemann sphere via the shape of its convex hull in hyperbolic 3-space. The authors ask whether finite width characterizes quasicircles—the curves that arise as limit sets of quasifuchsian groups and have controlled geometry. They show it does not: there exist Jordan curves with finite width that are not quasicircles (Theorem A). They then prove a positive result: a curve whose boundary width is smaller than the explicit constant $w_0 = \\cosh^{-1}(\\sqrt{2})$ is always a quasicircle, with a quantitative bound on its quasisymmetric constant (Theorems B and 3.1). The contrast with anti-de Sitter geometry, where finite width alone does characterize the analogous curves, is one of the paper's main points.","feed_headline":"Small width forces quasicircles; finite width alone does not","feed_subtitle":"Finite convex-hull width is not enough; the threshold is an explicit constant, cosh⁻¹(√2).","key_machinery":"The load-bearing objects are the convex hull $\\mathrm{CH}(C) \\subset \\mathbb{H}^3$ of a Jordan curve $C \\subset \\mathbb{CP}^1$, its two boundary disks $\\partial_+ \\mathrm{CH}(C)$ and $\\partial_- \\mathrm{CH}(C)$, and two invariants: the width $w(C) = \\sup_{x \\in \\mathrm{CH}(C)} (d(x,\\partial_+ \\mathrm{CH}(C)) + d(x,\\partial_- \\mathrm{CH}(C)))$ and the boundary width $w_\\partial(C) = \\max(\\sup_{x \\in \\partial_+ \\mathrm{CH}(C)} d(x,\\partial_- \\mathrm{CH}(C)), \\sup_{x \\in \\partial_- \\mathrm{CH}(C)} d(x,\\partial_+ \\mathrm{CH}(C)))$. The proof of the small-width theorem turns on a sharp distance estimate (Lemma 3.2): three hyperbolic planes bounding pairwise disjoint half-spaces force any point on one plane to be at distance at least $w_0 = \\cosh^{-1}(\\sqrt{2})$ from any point on another unless the geometry collapses; this constant emerges from a symmetric ideal-triangle configuration. The nearest-point projection $\\pi_+$ between the two boundary components provides the link to quasiconformality: Theorem C shows that $C$ is a quasicircle exactly when $\\pi_+$ is a quasi-isometry, and the small-width bound is used to prove that $\\pi_+$ is indeed a quasi-isometry.","core_discovery":"The paper's central discovery is that the relation between convex-hull width and quasicircles is more subtle in hyperbolic geometry than in its anti-de Sitter analogue. In the anti-de Sitter setting, the analogous width characterizes the analogous curves exactly. Here the authors define two hyperbolic invariants, the width $w(C)$ and the boundary width $w_\\partial(C)$, and prove that finiteness of $w(C)$ is necessary but not sufficient for $C$ to be a quasicircle; explicit curves with bounded width fail to be quasicircles (Theorem A). Conversely, the boundary width version is sufficient when small: if $w_\\partial(C) < w_0 = \\cosh^{-1}(\\sqrt{2})$, then $C$ is a quasicircle, with an explicit function $k(w)$ controlling the quasisymmetric constant for $w_\\partial(C) \\le w < w_0$ (Theorem 3.1). They also characterize quasicircles entirely in terms of the geometry of the convex hull: $C$ is a quasicircle if and only if a nearest-point projection $\\pi_+ \\colon \\partial_+ \\mathrm{CH}(C) \\to \\partial_- \\mathrm{CH}(C)$ is a quasi-isometry (Theorem C).","pith_inferences":["The threshold $w_0 = \\cosh^{-1}(\\sqrt{2}) \\approx 0.881$ is plausibly not optimal: the paper's own example only rules out the larger value $\\sinh^{-1}(\\sqrt{2}) \\approx 1.146$, leaving a gap; one could seek the true optimal threshold by optimizing over ideal-quadrilateral configurations rather than the symmetric case.","The quantitative version Theorem 3.1 suggests a possible bridge to minimal surface theory: following the analogy in Section 1.3, one might conjecture that a quasicircle of sufficiently small boundary width bounds a minimal surface in $\\mathbb{H}^3$ with principal curvatures $<1$.","The quasi-isometry criterion Theorem C could be tested numerically: given a Jordan curve's convex hull boundary, compute the Lipschitz constants of nearest-point projection; the curve should be a quasicircle exactly when those constants are finite, suggesting a computable quasiconformality estimator.","The construction of the non-quasicircle of finite width (Theorem A) relies on infinitely many neck-pinching scales; one might ask whether a single scale of pinching gives a finite-width non-quasicircle, which would simplify the example."],"forward_implications":["If a Jordan curve has boundary width below $w_0$, it is a quasicircle, and the quasisymmetric constant is controlled by an explicit function of the width.","There are Jordan curves with finite width that are not quasicircles, so any characterization of quasicircles by width must use a stronger condition than finiteness, such as smallness of boundary width.","The nearest-point projection criterion (Theorem C) gives a new geometric characterization of quasicircles in terms of the quasi-isometry type of the convex hull boundary, with constants mutually bounded by the quasicircle constant.","The width and boundary width are genuinely different invariants: there are sequences of quasicircles with uniformly bounded boundary width whose width tends to infinity (Proposition 4.1).","In the anti-de Sitter setting the analogous width characterizes quasicircles, while in hyperbolic space it does not; the paper delineates exactly what remains true."],"supporting_citations":[{"why":"Provides the AdS analogue (an acausal meridian is an Einstein quasicircle iff width < pi/2) that motivates the hyperbolic question and supplies the template for the width definition.","marker":"[BS10]"},{"why":"Gives the bounded-turning criterion for quasicircles that the paper uses to verify the constructed curves are or are not quasicircles.","marker":"[Ahl66]"},{"why":"Supplies the compactness of uniform quasicircles used in Lemma 2.2 to detect non-quasicircle limits.","marker":"[LV73]"},{"why":"Gives the extension of quasiconformal maps of CP^1 to bilipschitz maps of H^3, used in Theorem C to turn quasicircle regularity into width and quasi-isometry bounds.","marker":"[TV82]"},{"why":"Provides the bilipschitz identifications between the convex hull boundary components and the complementary domains, used in the converse direction of Theorem C.","marker":"[Sul81, EM86]"},{"why":"The Dehn-filling construction in this reference is used in Proposition 4.1 to build quasicircles with bounded boundary width but unbounded width.","marker":"[KT90]"},{"why":"Standard quasiconformal reflection arguments from this reference are used in the final step of Theorem C to conclude the curve is a quasicircle.","marker":"[Ahl63]"}],"fun_headline_variants":["Boundary width below cosh⁻¹(√2) forces quasicircles","Finite width insufficient; small boundary width suffices","Quasicircles need boundary width under cosh⁻¹(√2) — not just finite","In CP^1, small boundary width ⇒ quasicircle; finite width ⇏","Explicit width constant separates quasicircles from mere finiteness"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof of the key distance bound assumes, without a complete derivation, that the worst possible configuration is a symmetric ideal triangle with the point y at the axis of symmetry; if some other configuration of three disjoint half-spaces gave a smaller critical distance, the threshold $w_0$ and Theorem B would not hold.","fun_headline_variants_meta":{"raw":{"variants":["Boundary width below cosh⁻¹(√2) forces quasicircles","Finite width insufficient; small boundary width suffices","Quasicircles need boundary width under cosh⁻¹(√2) — not just finite","In CP^1, small boundary width ⇒ quasicircle; finite width ⇏","Explicit width constant separates quasicircles from mere finiteness"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001473,"raw_usage":{"total_tokens":5913,"prompt_tokens":928,"completion_tokens":4985,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":544,"completion_tokens_details":{"reasoning_tokens":4882}},"tokens_in":544,"tokens_out":4985,"duration_ms":35897,"temperature":1.0,"reasoning_tokens":4882,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T11:19:19.649568+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Exhibit a Jordan curve that is not a quasicircle and whose boundary width is strictly less than $\\cosh^{-1}(\\sqrt{2})$; Theorem B forbids it. A more targeted check: numerically optimize over triples of disjoint half-spaces in $\\mathbb{H}^3$ the quantity $\\max(d(x,y),d(x',y))$ as in Lemma 3.2; a value below $w_0$ would break the proof of Theorem B.","supporting_citations":[],"review_version":1}