{"id":"2d624946-1dea-47d9-9f84-054ba88979e6","arxiv_id":"2607.18989","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"Smooth stable families of Higgs bundles admit globally smooth normalized harmonic metrics, while polystable families can fail even continuously and require a new weak C0 harmonic mediator.","lead":"Stable families of Higgs bundles on compact Kähler manifolds are shown to admit global smooth harmonic metrics that vary smoothly with every parameter, and explicit counterexamples show that polystable families can fail even to admit continuous ones. The paper develops the deformation theory of the transform and a weaker operator-level correspondence that survives the failures.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Flatness upgrade in Theorem 1.1 depends on an unstated Chern–Weil identity: condition (9) may not by itself be the correct characteristic-number condition unless a Hodge-index/Bogomolov step is supplied.","rationale":"The reader's weakest assumption was precisely the flatness upgrade from a moment-map solution to a harmonic metric, resting on the numerical conditions and the Chern–Weil identity. I agree and sharpen that concern: the paper writes the identity only as a reference and does not demonstrate that (9), rather than the trace-free discriminant, is the exact vanishing condition. This is not a claim that the paper is wrong—the standard identity together with stability and the Bogomolov inequality may well fill the gap—but because every downstream stack-theoretic and infinitesimal statement inherits Theorem 1.1's 'harmonic' conclusion, and because Remark 5.3 explicitly isolates ν2 = 0 as the final topological input, an independent verification is warranted. I do not escalate to REJECT: the auditable analytic core is internally consistent, and the concern is a missing or under-specified standard identity, not a visible derivation error. Keeping the verdict CONDITIONAL reflects that the central claim should be accepted only after this check. The unverified negative examples in §14.3 are also a concern, but they are secondary to the stable theorem and would not affect the core construction if they failed; the numerical-condition identity is the most load-bearing point.","tokens_in":53043,"tokens_out":35016,"duration_ms":316630,"concrete_test":"Derive the Chern–Weil identity in the normalization of §2.1–§2.2 for a stable Higgs bundle on a compact Kähler surface (n=2). Express ||F_D + [θ,θ†]||^2_L2 as an explicit linear combination of ν1(E), ν2(E), and ∫c1(E)^2∧ω^{n-2}. Then check whether ν1=ν2=0 together with the Hodge-index theorem and the Bogomolov inequality forces the combination to vanish. As a concrete numerical probe, take a rank-2 stable Higgs bundle with θ=0 and c1=0, c2=1: the identity should give a nonzero norm and ν2≠0. Also test a candidate with c1·ω=0 and ch2=0; either no such stable bundle exists, or the identity should show zero curvature. If any nonzero curvature survives with ν1=ν2=0, Theorem 1.1 is false.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The load-bearing step is the promotion from a Hermitian–Yang–Mills–Higgs solution to a flat harmonic metric in §5.2/§5.3. Theorem 5.2 solves only the trace-free moment-map equation (10); determinant normalization gives the full moment-map equation μ=0. The proof then invokes 'the standard Hitchin–Simpson Chern–Weil identity' (Remark 5.3) to conclude flatness under (9). This is the only place the numerical conditions enter, yet the identity is never written in the paper's conventions. In the classical identity, the L2 norm of F_D + [θ,θ†] is controlled by a characteristic number that, when ν1=0, contains both ∫ch2∧ω^{n-2} and a ∫c1^2∧ω^{n-2} term. Condition (9) sets only the first to zero. Flatness follows only if stability/Bogomolov plus the Hodge-index theorem forces the second term to vanish. The paper does not supply that argument. If the identity or the ch2 convention in (8)–(9) is misstated, the Banach-IFT machinery still produces smooth moment-map solutions, but they are not flat; then Theorem 1.1's 'harmonic' conclusion, Corollary 1.2, Theorem 6.4, Theorem 1.3, and Theorem 1.4 fail as stated even though the analytic construction is untouched.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper develops the analytic foundations of the authors' earlier diffeological approach to non-Abelian Hodge theory. It proves that smooth families of stable Higgs bundles over arbitrary finite-dimensional parameter manifolds admit globally smooth harmonic metrics, provided each fibre satisfies the numerical conditions (9). The proof proceeds by local trivialization, Sobolev completions, a Banach implicit-function theorem for the trace-free moment-map equation, determinant normalization, and gluing by uniqueness. It then derives the first-variation formula L_h s_0 = -S_h(eta), identifies the differential of the transform, treats locally split polystable families of constant type, introduces relative harmonic filtrations with an obstruction theory, extends the results to finite regularity and reduced singular parameter spaces, discusses heat-flow limits, and constructs a smooth Hodge lambda-family on the stable locus. The negative examples in Section 14.3 are used to define a weak C0 operator-level harmonic mediator.","tokens_in":53265,"tokens_out":19032,"duration_ms":174923,"significance":"If the analytic core is correct, the paper fills a substantial gap in the diffeological framework: it proves that stable families lie in the essential image of the harmonic mediator before extension completion, and it gives a plotwise smooth non-Abelian Hodge transform. The fixed-Banach-space formulation (eqs. 24-25), the identification of the Jacobi operator as (D'')*D'' (Prop. 4.2), the stability-to-simplicity argument (Prop. 4.4), the determinant Poisson step (Prop. 5.1), the normalized uniqueness via Donaldson-functional convexity (Prop. 5.5), and the gluing argument (Cor. 5.6) are clearly presented and auditable. The paper is honest in separating the moment-map problem from the flatness upgrade, and the explicit examples of polystable families without continuous harmonic metrics, if correct, are a strong and falsifiable contribution. The first-variation and Green-operator formulas provide useful tools for deformation theory.","major_comments":[{"comment":"The flatness upgrade is not proved. The text invokes a 'standard Hitchin–Simpson Chern–Weil identity' but never writes it. The relevant characteristic-number combination is the Bogomolov-type expression ∫(2r c2−(r−1)c1^2)∧ω^{n−2}, not ∫ch2∧ω^{n−2} alone. As written, (9) sets only ν1,ν2 to zero; the paper does not supply the Hodge-index/Bogomolov argument that this forces the curvature energy to vanish. Since Theorem 1.1's harmonic conclusion and all subsequent plotwise/stack results depend on this step, the proof is incomplete. Please state the identity in the paper's conventions and either prove the implication from (9) for stable fibres or strengthen (9) to full Chern-class vanishing.","section":"§5.2–5.3, Remark 5.3, eqs. (8)–(9)"}],"minor_comments":[{"comment":"The definition of the trace-free target T^0_k depends on the fixed background h0, while the source uses the moving metric. This is clarified by the isometry I_s, but a one-sentence reminder before (24) would help readers.","section":"§3.1, eq. (24)"},{"comment":"The phrase 'the standard Chern–Weil identity expresses the remaining curvature energy ... as the characteristic-number combination in (9)' is imprecise: the identity involves a linear combination of c1 and c2 (or ch2), not the two separate numbers ν1 and ν2. This is the same point as the major comment; please rewrite this sentence after adding the explicit identity.","section":"§5.3, proof of Theorem 1.1"},{"comment":"The definition of mixed regularity would be clearer if phrased as: in local trivializations, the coefficient map u ↦ coefficient(u,·) is C^d into the Fréchet space C^∞(X), rather than saying the coefficients themselves are 'smooth in X and C^d in u'.","section":"§11.1, Definition 11.1"},{"comment":"For the endpoint j=0, the use of W^{-1,2} and elliptic duality is invoked without a precise definition or reference. Adding a short explanation or a citation for the negative-order Sobolev spaces would improve readability.","section":"§7.4, Proposition 7.11"}],"recommendation":"major_revision","confidential_remarks":"The only obstacle I see to eventual acceptance is the missing Chern–Weil identity in the flatness step. This is load-bearing but readily fixable by adding a lemma and either proving the Hodge-index/Bogomolov implication from (9) or adopting the stronger Chern-class vanishing hypothesis. I would also ask a second referee to check the Section 14.3 examples carefully, since they support the paper's negative claims and the weak C0 mediator."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nYou should know this is a real sequel with a substantial analytic theorem, and the arguments I could audit hold together. The headline result is a global smooth harmonic metric for any smooth stable Higgs family over an arbitrary finite-dimensional plot, with the first-variation Green-operator formulas and a smooth diffeological transform on the stable locus. That is genuinely new in its plotwise form, and the paper is honest that the local analytic dependence was already known from Hu–Sun–Yang–Zuo and Kim–Wilkin; the novelty is the globalization, the stack-level formulation, and the deformation theory.\n\nThe analytic core is standard but carefully done. The Jacobi operator identity, stability-to-simplicity, the fixed-Banach-space IFT, and the gluing are all consistent as far as I could audit. The constant-type polystable theorem is a nice clean statement.\n\nOne worry circulated in our stress test was that the flatness upgrade from a moment-map solution needs a Chern-Weil identity that isn't written. I don't think this worry lands. They define ν2 via ch2, and the standard Higgs curvature energy identity says the L2 norm of the primitive curvature is proportional to the integral of ch2 wedge ω^{n-2}. With μ=0, that's exactly the identity; no extra Hodge-index step is required. The paper's shorthand \"standard Chern-Weil identity\" is terse, but the logic is sound. A referee should still ask them to write the identity explicitly, because it's the only place the numerical conditions enter.\n\nThe real soft spot is the negative material in §14.3–14.4. The abstract and introduction make strong claims: real-analytic polystable families with no continuous harmonic metric, no relative harmonic filtration, and the weak C0 mediator with a stack equivalence. Those are the headline news for the polystable side, but they're in the final section, and I couldn't verify them in detail. The rest of the paper doesn't depend on them, so the stable and constant-type results stand regardless. If the examples are right, the paper is a significant contribution to the family-level non-abelian Hodge story; if they're wrong, the polystable claims collapse but the stable part survives.\n\nThe main dependency is on the prior paper's diffeological stack framework. If you don't buy that framework, the stack-level consequences won't matter much to you, but the analytic theorems are independent.\n\nVerdict: This deserves a serious referee. It's long and could be tightened, and the examples need careful checking, but the core is original and sound. I'd send it to review.","headline":"Solid analytic core for stable families, with genuinely new plotwise smoothness and variation formulas; the negative polystable examples are the part to scrutinize.","tokens_in":53933,"tokens_out":13141,"would_cite":true,"duration_ms":114646,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper proves that for every smooth family of stable Higgs bundles on a compact Kähler manifold satisfying the standard numerical conditions, a harmonic metric can be chosen smoothly in the family parameter, making the non-Abelian Hodge","keywords":["diffeological stacks","harmonic metrics","Higgs bundles","non-Abelian Hodge correspondence","relative deformation theory","Sobolev completions","polystable families","Hitchin–Simpson equation"],"falsifier":"Search for a smooth family of stable Higgs bundles over a compact parameter manifold, satisfying the numerical conditions, whose normalized fibrewise harmonic metric is not smooth in the parameter. Theorem 1.1 asserts that no such family exists; a counterexample (constructed, say, by inducing a harmonic metric through a family of gauge transformations whose smoothness degenerates) would refute the central claim. Alternatively, for a concrete family with all assumptions satisfied, compute the curvature tensor of the associated connection and verify the Chern–Weil identity; if the curvature ener","tokens_in":52697,"feed_emoji":"","tokens_out":6473,"duration_ms":59318,"temperature":0.7,"pith_summary":"The paper establishes that the fibrewise harmonic metric of a stable Higgs bundle varies smoothly in the parameter of a smooth family, provided the usual non-Abelian Hodge numerical conditions hold. The proof runs the Hitchin–Simpson moment-map equation on Sobolev completions over arbitrary finite-dimensional parameter plots, uses stability to make the linearized Jacobi operator invertible on a trace-free slice, and glues local solutions via a determinant normalization. From this existence-plus-regularity result the paper derives that the non-Abelian Hodge transform is smooth along diffeological plots and computes its first variation explicitly via a Green operator. On the polystable locus the paper shows the analogous statement fails: real-analytic families can lack continuous harmonic metrics, so the correspondence is preserved only through a larger weak C^0 operator-level mediator. For semistable families, the extension-generated stack is characterized by relative harmonic filtrations, with an obstruction theory for assembling them.","feed_headline":"Stable Higgs families admit smooth harmonic metrics","feed_subtitle":"The non-Abelian Hodge transform becomes a smooth morphism of diffeological stacks on the stable locus, not just a pointwise correspondence.","key_machinery":"The load-bearing machinery is the parameter-dependent Banach-space formulation of the Hitchin–Simpson moment map. A family of Higgs bundles is viewed as a smooth map into an ambient Sobolev affine space; the harmonic equation becomes a nonlinear map F(u,s) on fixed Sobolev spaces, where s is the logarithmic metric variation in an exponential gauge. The metric-direction derivative is the Jacobi operator L_h = (D'')*_h D'' (the Higgs Laplacian on Hermitian endomorphisms), which is self-adjoint, elliptic, and nonnegative. On a stable Higgs bundle its kernel consists of scalar Hermitian Higgs endomorphisms, so after imposing a determinant normalization and projecting to the trace-free slice, L_h","core_discovery":"The central discovery is that the non-Abelian Hodge correspondence, which classically relates stable Higgs bundles to irreducible flat connections fibrewise, is itself a smooth operation on families: given a smooth family (E, D'') of stable Higgs bundles over a compact Kähler manifold X parametrized by a smooth manifold U, with all fibres satisfying the numerical conditions ν1=ν2=0, there exists a global smooth Hermitian metric h on E such that h_u is harmonic for each fibre. After fixing a Hermitian–Einstein determinant metric, the harmonic metric is unique, so local solutions glue globally. The proof converts the Hitchin–Simpson equation into a nonlinear map on Sobolev spaces whose metric","pith_inferences":["The analytic machinery is likely robust enough to handle families with additional structure (e.g., parabolic or twisted Higgs bundles) where the same Banach-space implicit-function argument would yield analogous smooth dependence, though the flatness upgrade would need the appropriate numerical conditions.","The spectral-gap perspective in the paper suggests a quantitative version: the norm of the first variation of the harmonic metric near a polystable degeneration is controlled by the inverse spectral gap, indicating that smoothness degrades in a predictable way as the stabilizer grows; this could be made into a testable regularity estimate.","The weak C^0 operator-level mediator, although not a smooth transform, might be a natural object for studying families of semisimple local systems that cross separatrices or non-closed orbits, where the usual harmonic metric degenerates but the adjoint operators still converge.","If the numerical conditions are relaxed, the paper's proof still yields smooth solutions to the moment-map equation (not necessarily flat); these could be interpreted as 'almost harmonic' metrics and might be used to construct approximate Hodge systems with controlled error, quantifying the role of the Chern-class conditions."],"forward_implications":["The non-Abelian Hodge transform is a smooth morphism of diffeological stacks on the stable locus, with smooth inverse on irreducible families, so the classical correspondence holds at the level of smooth families, not just pointwise.","The first variation formula gives a concrete Green-operator expression for the infinitesimal transform along any plot, agreeing with the classical cohomological comparison at unobstructed points.","Stable loci lie in the harmonic-image substack before any extension completion or stackification, answering the stable case of the earlier open question about smooth parameter dependence.","The smooth Hodge λ-family on the stable locus supplies a finite-dimensional-parameter family of λ-connections, with pullback functoriality, giving a diffeological Hodge enhancement.","The paper's negative examples for polystable families and the weak C^0 mediator show that extension completion is genuinely necessary outside the stable/constant-type locus, and that the correspondence can persist in a weaker operator-level sense even when continuous harmonic reductions fail."],"fun_headline_variants":["Stable Higgs families now have smooth harmonic metrics","Hodge correspondence is smooth on diffeological stacks","Smooth harmonic metrics for stable Higgs family deformations","Family-wise smooth Hodge theory proved via Sobolev spaces","Non-Abelian Hodge: smooth fibrewise harmonic metrics"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The theorem's flatness conclusion depends on the exact numerical conditions ν1(E)=0 and ν2(E)=0 holding for every fibre; the analytic construction only solves the trace-free moment-map equation, and without those Chern-Weil identities the resulting Hermitian–Einstein–Higgs metric need not be flat, so 'harmonic' in the paper's strong sense would fail even though the smooth family of metrics exists.","fun_headline_variants_meta":{"raw":{"variants":["Stable Higgs families now have smooth harmonic metrics","Hodge correspondence is smooth on diffeological stacks","Smooth harmonic metrics for stable Higgs family deformations","Family-wise smooth Hodge theory proved via Sobolev spaces","Non-Abelian Hodge: smooth fibrewise harmonic metrics"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000256,"raw_usage":{"total_tokens":1490,"prompt_tokens":901,"completion_tokens":589,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":645,"completion_tokens_details":{"reasoning_tokens":509}},"tokens_in":645,"tokens_out":589,"duration_ms":5335,"temperature":1.0,"reasoning_tokens":509,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-01T13:50:33.444083+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Search for a smooth family of stable Higgs bundles over a compact parameter manifold, satisfying the numerical conditions, whose normalized fibrewise harmonic metric is not smooth in the parameter. Theorem 1.1 asserts that no such family exists; a counterexample (constructed, say, by inducing a harmonic metric through a family of gauge transformations whose smoothness degenerates) would refute the central claim. Alternatively, for a concrete family with all assumptions satisfied, compute the curvature tensor of the associated connection and verify the Chern–Weil identity; if the curvature ener","supporting_citations":[],"review_version":1}