{"id":"0313efe1-6145-4037-83ba-e46ec1994ed6","arxiv_id":"2505.01943","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Rough singular Kähler-Einstein metrics, when the metric completion is an RCD space, are equivalent to the Eyssidieux-Guedj-Zeriahi notion and imply log terminal singularities; this applies to tangent cones of noncollapsed Kähler limits.","lead":"This paper proves that singular Kähler-Einstein metrics with bounded potentials and good metric limits (RCD spaces) force the underlying variety to have log terminal singularities, and that two common definitions of singular Kähler-Einstein metric coincide. This settles a conjecture by Song Sun and a question by Hallgren on limits of Kähler-Einstein metrics and Kähler-Ricci flows.","discovery_kind":"unification","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Proposition 4.2's proof of RCD for Kähler-Ricci cone limits invokes an unproven Sobolev-to-Lipschitz property; without it, Theorem 1.6 for flow limits is unsupported.","rationale":"The reader's weakest_assumption correctly identifies the RCD condition and epsilon-regularity as load-bearing, but stops short of noticing that in the Kähler-Ricci flow application the RCD condition is itself established by a highly compressed argument. Proposition 4.2 is the sole justification for Definition 1.2(iv) in case (B), and its proof appears to require an unstated Sobolev-to-Lipschitz property. The subsequent verification of Lipschitz regularity for harmonic functions is insufficient if Honda's theorem needs the stronger property for all W^{1,2} functions. I do not claim the result is false; the gap is a missing verification, not a known contradiction. However, because the paper marks Theorem 1.6 for Kähler-Ricci flows as a main application, the proof should be completed or cited precisely before the verdict stands. The recommended verdict is therefore CONDITIONAL rather than REJECT: accept if the Sobolev-to-Lipschitz property can be verified from the cited ingredients, otherwise the flow-limit theorem is unproven.","tokens_in":20856,"tokens_out":28367,"duration_ms":289257,"concrete_test":"Read [Hon18, Corollary 3.10] and list all hypotheses. Then prove the Sobolev-to-Lipschitz property for C(Z) directly from the heat-kernel bounds in [Hal24, Appendix] and the Sobolev inequality in [CMZ24, Cor 1.9], without assuming Z is RCD. If the proof cannot be completed, test the statement on a concrete cone C(Z) that satisfies the other stated assumptions: if such a cone admits f in W^{1,2}(C(Z)) with |∇f| ≤ 1 but no Lipschitz representative, Proposition 4.2 is false.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"Definition 1.2(iv) (RCD) is the entry point for all heat-kernel and Moser arguments, and for case (B) (Kähler-Ricci flow limits) it is established only in Proposition 4.2. That proof is one paragraph: it reduces to Honda's characterization [Hon18, Cor 3.10] and asserts that C(Z) has 'the Sobolev to Lipschitz property', with no citation or derivation. The next two sentences establish a Sobolev inequality via [CMZ24, Cor 1.9] and then verify that eigenfunctions v in W^{1,2}(Z) with Δv = -λv are Lipschitz via Lemma 4.1. Neither of these is equivalent to Sobolev-to-Lipschitz for arbitrary W^{1,2} functions. If [Hon18, Cor 3.10] really requires this property, then the proof has a gap: one cannot conclude C(Z) is RCD(0,2n), and hence Theorem 4.3 (and Theorem 1.6) for case (B) does not follow from Theorem 1.5. This is load-bearing because the Kähler-Ricci flow application is a key novelty beyond case (A).","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper introduces the notion of a rough Kähler–Einstein variety (Definition 1.2): a normal Kähler variety with a smooth Kähler–Einstein metric on the regular set, bounded local potentials, local domination, an RCD(λ,2n) condition on the metric completion, and an ε-regularity condition. The main theorem (Theorem 1.3) states that every such variety has log-terminal analytic germs; for compact or conical rough Kähler–Einstein varieties, Theorem 1.5 adds global Q-Gorensteinness, existence of the EGZ singular Kähler–Einstein metric, algebraic volume ratio, and uniqueness of the Ricci-flat cone. Theorem 1.6 applies these results to Ricci-flat cones arising as noncollapsed limits of Kähler–Einstein manifolds or Kähler–Ricci flows, resolving a conjecture from [Sun25] and a question from [Hal24]. The proof combines RCD-space tools (heat-kernel estimates, cutoffs, Moser iteration, improved Kato inequalities) with the Donaldson–Sun construction of peaked holomorphic sections.","tokens_in":21135,"tokens_out":16431,"duration_ms":166021,"significance":"If correct, the paper gives a clean and fairly general bridge between the weak notion of a smooth Kähler–Einstein metric on the regular set with bounded potentials and the stronger EGZ notion, under an RCD hypothesis. The elliptic estimates in Section 2 are carefully written, and the adaptation of Donaldson–Sun to the non-Q-Gorenstein setting is a substantial technical achievement. The case of limits of Kähler–Einstein manifolds (case (A)) appears well supported. However, the advertised application to Kähler–Ricci flow limits (case (B)) rests on Proposition 4.2, whose proof contains an unsubstantiated Sobolev-to-Lipschitz assertion; this is a load-bearing gap. With that gap repaired, the results would be a significant advance.","major_comments":[{"comment":"The proof asserts that C(Z) satisfies the Sobolev-to-Lipschitz property with no citation or derivation. Honda's characterization [Hon18, Corollary 3.10] appears to require this property as a hypothesis, and the subsequent argument only proves that eigenfunctions of Δ_Z with fixed eigenvalue are Lipschitz, which does not imply Sobolev-to-Lipschitz for arbitrary W^{1,2} functions. Lemma 4.1 handles only dilation-homogeneous harmonic functions on the cone. Consequently the RCD(0,2n) conclusion for case (B) is not established as written, and therefore Theorem 4.3 and Theorem 1.6 do not follow for Kähler–Ricci flow limits. Please supply a proof or a precise reference for the Sobolev-to-Lipschitz property in this setting, or state it as an additional hypothesis if it is not available.","section":"Section 4, Proposition 4.2"},{"comment":"The proof of Theorem 1.5(i) jumps from the local construction of a bounded section to the assertion that X has log-terminal singularities via [EGZ09, Lemma 6.4], and the statement also asserts global Q-Gorensteinness. Log terminality and Q-Cartierness at each point do not by themselves imply that a fixed power of K_X is a line bundle on a compact X, as Remark 1.4 itself notes for noncompact X. Please either cite the standard boundedness of the Cartier index for klt singularities in fixed dimension or provide the compactness argument; otherwise Theorem 1.5(i) is not fully proved.","section":"Section 3, Theorem 1.5(i)"}],"minor_comments":[{"comment":"The volume threshold in (v) uses H^{2n}(B(x,r)) ≥ (ω_{2n}−ε)r^{2n}; for consistency with the RCD measure in (iv), please specify explicitly that H^{2n} is the Hausdorff measure of the metric completion and that ω_{2n} is the Euclidean unit-ball volume.","section":"Definition 1.2"},{"comment":"In the construction of α, the case λ=0 gives α=0, contradicting the displayed 'α>0'. Since eigenvalue-zero eigenfunctions are constant on a connected cross-section, this is harmless but should be noted.","section":"Proposition 4.2"},{"comment":"The heat-kernel time variable is written as 1−t throughout the proof, while earlier sections use r²−t for a scale r. The notational mismatch should be harmonized to avoid confusion.","section":"Lemma 4.1"},{"comment":"The phrase 'argue as in [DS14, Section 3.2.2]' covers a substantial part of the construction of the peaked section v. Since the current setting is not Q-Gorenstein, a slightly longer explanation of which steps of [DS14] carry over verbatim would improve readability.","section":"Theorem 1.3"}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"You should know this paper for the C0 and C1 estimates in Section 2, which are a genuinely new tool for constructing sections of powers of the canonical bundle on singular Kähler-Einstein spaces without assuming Q-Gorenstein. The main theorem—rough Kähler-Einstein varieties are log terminal—resolves a conjecture of Sun and a question of Hallgren, and the reduction to the Donaldson–Sun method via RCD heat kernel bounds is well executed. The core is sound.\n\nThe soft spot is Proposition 4.2. Theorem 1.6 for Ricci-flow limits rests on proving that the cone C(Z) from an F-limit is RCD(0,2n). The one-paragraph proof invokes Honda's characterization and asserts that C(Z) has the Sobolev-to-Lipschitz property, with no citation or argument. It then verifies only that eigenfunctions (homogeneous harmonic functions) are Lipschitz via Lemma 4.1, which is weaker than Sobolev-to-Lipschitz for arbitrary W^{1,2} functions. Unless Honda's Corollary 3.10 does not require that input—which the text suggests it does—this is a genuine gap: Theorem 1.6 case (B) does not follow from the stated proof. It is likely fixable, by combining the Sobolev inequality cited from [CMZ24] with the doubling and Poincaré structure of the cone, or by locating the property in [Bam21] or [Hal24], but as written it is unsupported. This does not affect Theorems 1.3 and 1.5, and the KE-limit case (A) is fine because RCD stability under GH convergence applies directly.\n\nThe paper deserves a serious referee. The main ideas are new, the writing is honest (they mention the independent SWZ25 preprint), and the gap is localized. I would accept for peer review and push for a strengthened proof of Proposition 4.2. I would cite the paper for the rough KE result and the elliptic estimates; the flow-limit conclusion I would cite with a caveat until the gap is closed.","headline":"Strong paper, but the RCD proof for flow-limit cones has a real gap in the Sobolev-to-Lipschitz step.","tokens_in":21671,"tokens_out":13623,"would_cite":true,"duration_ms":144182,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["32Q20","53C23","53C55"],"pacs":[],"model":"deepseek-v4-flash","headline":"Under a synthetic Ricci bound, every rough Kähler-Einstein variety has log terminal singularities.","keywords":["singular Kähler-Einstein metrics","RCD spaces","log terminal singularities","Q-Gorenstein","Kähler-Einstein cones","Kähler-Ricci flow","noncollapsed limits","Monge-Ampère equation"],"falsifier":"Exhibit a rough Kähler-Einstein variety (smooth Kähler-Einstein metric on the regular locus, bounded local potentials, RCD completion, epsilon-regularity) at some point of which the canonical bundle has infinite index. The theorem asserts no such point exists; finding one would refute the central claim. A practical place to look would be non-log-terminal Kähler cones with Ricci-flat cone metrics, checking whether their metric completions are RCD.","tokens_in":20663,"feed_emoji":"📐","tokens_out":12505,"duration_ms":119177,"temperature":0.7,"pith_summary":"The paper asks whether a smooth Kähler-Einstein metric on the regular part of a normal variety, with bounded local potentials, forces the variety to have log terminal singularities. It answers yes whenever the metric completion satisfies a synthetic Ricci lower bound (the RCD condition) together with a quantitative epsilon-regularity condition; such spaces are called rough Kähler-Einstein varieties. The main theorem shows every analytic germ of a rough Kähler-Einstein variety is log terminal, and in the compact or Ricci-flat cone case the metric extends to a singular Kähler-Einstein metric in the Monge-Ampère sense, with algebraic volume ratio in the cone case. Since noncollapsed limits of Kähler-Einstein manifolds and Kähler-Ricci flows automatically meet the rough Kähler-Einstein hypotheses, the two competing notions of singular Kähler-Einstein metric coincide in precisely the settings where these limits appear.","feed_headline":"Weak singular Kähler-Einstein metrics force log terminal singularities","feed_subtitle":"A synthetic Ricci bound makes the weak definition of singular Kähler-Einstein metric as strong as the standard one.","key_machinery":"The central object is the rough Kähler-Einstein variety, defined by a smooth Kähler-Einstein metric on the regular locus with bounded local potentials, local domination of a smooth metric, an RCD metric completion (a synthetic Ricci lower bound together with a dimension bound), and an epsilon-regularity condition. The proof's workhorse is a package of analytic estimates: improved Kato inequalities and Gaussian heat-kernel bounds on the RCD space yield $C^{0}$ and $C^{1}$ control for holomorphic sections of powers of the canonical bundle on the regular locus, after showing the singular set has Hausdorff codimension at least four. L2 estimates for the ∂-operator turn almost-holomorphic peaked sections into genuine holomorphic sections. The peaked sections, concentrated near a prescribed point, imply that a fixed power of the canonical bundle extends across that point, which is exactly the Q-Gorenstein and log-terminal conclusion.","core_discovery":"The central claim is that the weak and strong notions of singular Kähler-Einstein metric are equivalent under a synthetic Ricci bound. A rough Kähler-Einstein variety, defined by a smooth Kähler-Einstein metric on its regular locus, bounded local potentials, local metric domination, an RCD completion, and an epsilon-regularity condition, has log terminal singularities at every point. In the compact or Ricci-flat cone cases this strengthens to: the variety is Q-Gorenstein, the metric extends to a Kähler current solving the Monge-Ampère equation, and in the cone case the volume ratio is algebraic and the cone is the unique Ricci-flat Kähler cone with its Reeb vector field. Consequently every Ricci-flat metric cone arising as a noncollapsed limit of Kähler-Einstein manifolds or Kähler-Ricci flows satisfies these conclusions.","pith_inferences":["The RCD condition is likely the right synthetic hypothesis for this circle of ideas: it packages the analytic input needed for the section estimates, and the paper suggests that relaxing it would require a wholly different method rather than a small tweak.","The same C^0 and C^1 section estimates could be applied to other Hermitian holomorphic line bundles on singular Kähler spaces, potentially yielding pluricanonical extension theorems beyond the Kähler-Einstein setting.","A testable extension is whether the epsilon-regularity condition is redundant: if it follows from the RCD condition together with bounded potentials, the definition of rough Kähler-Einstein variety could be simplified, widening the class of spaces covered."],"forward_implications":["Every Ricci-flat metric cone arising as a noncollapsed limit of Kähler-Einstein manifolds or Kähler-Ricci flows has log terminal singularities and is Q-Gorenstein.","On such spaces, any smooth Kähler-Einstein metric with bounded potentials on the regular locus extends to a singular Kähler-Einstein metric in the Monge-Ampère sense, so the weak and strong notions agree.","For such cones, the volume ratio is an algebraic number, and the cone is the unique Ricci-flat Kähler cone on its underlying variety with the given Reeb vector field.","The local nature of the result means every analytic germ at a singularity of a rough Kähler-Einstein variety is log terminal, not only in the compact or cone cases."],"supporting_citations":[{"why":"Defines the stronger singular Kähler-Einstein metric by a Monge-Ampère equation and provides the log-terminal criterion used to conclude the metric extension.","marker":"[EGZ09]"},{"why":"Supplies the method of peaked holomorphic sections for showing powers of the canonical bundle extend.","marker":"[DS14]"},{"why":"Provides the algebraic uniqueness and volume-ratio conclusions in the Ricci-flat cone case.","marker":"[DS17]"},{"why":"Establishes that tangent cones of noncollapsed RCD spaces are metric cones, the structure used to compare with flat space.","marker":"[DPG18]"},{"why":"Rules out boundary-like tangent cones, giving the codimension bound on the singular set.","marker":"[BNS22]"},{"why":"Provides local cutoff functions on RCD spaces used in the elliptic estimates.","marker":"[MN19]"},{"why":"Gives the Gaussian heat-kernel upper bounds on the RCD space that drive the C^0 and C^1 estimates.","marker":"[JLZ16]"},{"why":"Gives the complete Kähler metric and L2 ∂-estimates needed to perturb approximate sections to holomorphic sections on the regular locus.","marker":"[Dem82]"},{"why":"Provides the epsilon-regularity theorem for Einstein metrics used to verify the regularity condition and smooth convergence.","marker":"[And90]"},{"why":"Supplies the structure theory of noncollapsed Ricci-flow limits, including the entropy criterion used to verify the epsilon-regularity condition in the flow case.","marker":"[Bam21]"}],"fun_headline_variants":["Weak equals strong for singular KE metrics under Ricci bound","RCD bound turns weak singular KE metrics strong","Log terminal singularities forced by weak KE metrics with Ricci bound","Synthetic Ricci bound equates weak and strong singular KE metrics","On RCD spaces, weak singular KE metrics imply log terminal"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the metric completion of the regular locus, with its volume measure, satisfies the RCD condition (a synthetic Ricci lower bound with dimension control) together with the epsilon-regularity condition; if this premise fails, the equivalence of the two notions is not claimed.","fun_headline_variants_meta":{"raw":{"variants":["Weak equals strong for singular KE metrics under Ricci bound","RCD bound turns weak singular KE metrics strong","Log terminal singularities forced by weak KE metrics with Ricci bound","Synthetic Ricci bound equates weak and strong singular KE metrics","On RCD spaces, weak singular KE metrics imply log terminal"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.002181,"raw_usage":{"total_tokens":8394,"prompt_tokens":831,"completion_tokens":7563,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":447,"completion_tokens_details":{"reasoning_tokens":7482}},"tokens_in":447,"tokens_out":7563,"duration_ms":51117,"temperature":1.0,"reasoning_tokens":7482,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T04:05:27.951081+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Exhibit a rough Kähler-Einstein variety (smooth Kähler-Einstein metric on the regular locus, bounded local potentials, RCD completion, epsilon-regularity) at some point of which the canonical bundle has infinite index. The theorem asserts no such point exists; finding one would refute the central claim. A practical place to look would be non-log-terminal Kähler cones with Ricci-flat cone metrics, checking whether their metric completions are RCD.","supporting_citations":[],"review_version":1}