{"id":"2580bf18-9325-470c-b618-1cf3b2c5ccff","arxiv_id":"1908.11235","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A proof that toroidal crossing spaces with a simple logarithmic section and a transverse anticanonical divisor admit smoothings, together with a proof of Danilov's Hodge-de Rham degeneration conjecture.","lead":"This paper proves that certain singular spaces called toroidal crossing spaces can be smoothed into smooth or mildly singular manifolds when they carry a suitable logarithmic structure. It also settles a 1978 conjecture by Danilov on the degeneration of a Hodge-de Rham spectral sequence for toroidal pairs.","discovery_kind":"first_principles","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 1.10, and hence the Maurer-Cartan smoothing step, depends on Lemma 12.1, whose local description rests on the unproved analytic Lemma 7.14; this is the load-bearing soft spot, not the unproved arbitrary-coherent-base conjecture.","rationale":"The reader's verdict is CONDITIONAL and identifies the restriction to sharp-toric-monoid bases as the weakest assumption. I agree that the proof is local-model-based, but the unrestricted coherent-base conjecture is not needed for Theorem 1.7: the smoothing argument only invokes Theorem 1.10 over S_m, which is of the covered form. The genuinely load-bearing point is the internal solidity of the proof of Theorem 1.10, specifically Lemma 12.1. That lemma is the bridge between the algebraic ETD computations and the analytic acyclicity needed for Katz's base-change argument. Its proof depends on Lemma 7.14, whose proof is explicitly omitted, and on a lengthy analytic induction. If Lemma 7.14 is false or incomplete in the relevant case, the local description of the kernel complex K^bullet is wrong, the acyclicity assertion fails, and Theorem 1.10 cannot be used to prove the surjectivity in Theorem 13.1. This would remove the existence of the Maurer-Cartan solution and hence the formal smoothing. I do not see a concrete contradiction in the paper, and the overall strategy is coherent, so the concern does not force a stronger verdict than the reader's CONDITIONAL. The proposed check is designed to settle the one unproved analytic lemma on which the chain of reasoning rests.","tokens_in":40029,"tokens_out":41983,"duration_ms":418088,"concrete_test":"Recompute the stalk in Lemma 7.14 for a concrete ETD with one-dimensional base where two different V_e occur: take Q = N and P = N x P0 with P0 the monoid from Example 7.5, F = F_min, and S_1 = Spec(N -> C[t]/(t^2)). Compute both sides of Lemma 7.14 for the module tilde-V from Corollary 7.11 at the origin: the completed tensor product with O_{Y^an,0} versus the growth-condition submodule of V[[E_K]]. If they are not equal, Lemma 12.1's local model is invalid. A positive check would still be valuable: implement the descent induction in Lemma 12.1 for this ETD and verify H^0(K^bullet)_0 = H^1(K^bullet)_0 = 0.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central smoothing claim (Theorem 1.7) is obtained by solving the Maurer-Cartan equation in Section 13; the two inputs that make the formal deformation exist are Theorem 13.1 and Theorem 13.3, both derived from the one-parameter relative degeneration Theorem 1.10. The proof of Theorem 1.10 (Section 12.1) reduces to the statement that the kernel complex K^bullet is acyclic at the origin for every elementary log toroidal model with one-dimensional base: this is Lemma 12.1. Lemma 12.1 is not a routine verification: its local description of K^bullet uses Corollaries 7.11 and 7.12 together with Lemma 7.14, and the proof of Lemma 7.14 is explicitly left to the reader ('We leave the technical details to the reader'). The acyclicity argument then proceeds by a delicate analytic induction with norm estimates and two linear-algebra claims. If Lemma 7.14 fails for the relevant modules -- e.g. if the analytic tensor product of tilde-V with O_{Y^an,0} is strictly larger than the growth-condition completion -- then the formula for the stalk of W^bullet used throughout Lemma 12.1 is wrong, the acyclicity conclusion is unsupported, and Theorem 1.10 (freeness and E1-degeneration over S_m) would not be established. Since Theorem 13.1 uses Theorem 1.10 to prove the surjectivity that produces the Maurer-Cartan solution, the smoothing theorem would lose its formal-deformation step. The reader's identified limitation (the conjecture for arbitrary coherent bases) is not the real risk: Theorem 1.7 only needs the sharp-toric-monoid bases Spec(N -> k) and S_m, which are covered. The brittle point is the unproved analytic lemma used inside the covered case.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proves a smoothing theorem for proper toroidal crossing spaces under a 'simple section' hypothesis on the sheaf LS_X and an anticanonical section transverse to all relevant strata (Theorem 1.7), and derives the normal crossing case (Theorem 1.1). The proof equips X with a log toroidal family structure, studies the reflexive log de Rham complex W^•_{X/S}, proves Hodge–de Rham degeneration for such families (Theorem 1.9) and its relative one-parameter version (Theorem 1.10), and then uses Maurer–Cartan/Batalin–Vilkovisky methods to construct a formal smoothing, which is upgraded to an analytic smoothing via Grauert–Douady/Artin approximation. Along the way the paper settles Danilov's conjecture for proper toroidal pairs (Theorem 1.4). The local arguments are carried out in elementary log toroidal models, with explicit computations in Sections 7 and 12.","tokens_in":40346,"tokens_out":15099,"duration_ms":153871,"significance":"If correct, this is a substantial advance: it replaces the local rigidity assumption in Gross–Siebert smoothing with a much milder condition involving a simple section and generation of T^1_X, and it gives a general smoothing criterion for toroidal crossing spaces. The proof strategy, combining a Deligne–Illusie degeneration argument with BV/Maurer–Cartan deformation theory, is novel and likely to be influential. The paper is also notable for its explicit local toric computations (Propositions 7.2, 7.3, Corollaries 7.11, 7.12, Lemma 12.1) and for clearly stating the scope of Theorem 1.9, including the conjecture for arbitrary coherent bases. The main weakness is that one key analytic lemma, Lemma 7.14, is stated and then left to the reader; since Lemma 12.1 and hence Theorem 1.10 depend on it, this is a load-bearing gap that must be closed before the main smoothing claim is fully supported.","major_comments":[{"comment":"Lemma 7.14 is load-bearing but its proof is reduced to 'We leave the technical details to the reader.' The lemma identifies the stalk at the origin of V^an = ~V ⊗_{C[E_K]} O_{Y^an} with the growth-condition completion of V[[E_K]]. This identification is used in Remark 7.15 to describe W^{m,an}_{Y/T} and W^{m,an}_Y, and Lemma 12.1 relies on that description to prove the acyclicity of K^•. Since Theorem 1.10, and through it Theorem 13.1 and the formal deformation step, depends on Lemma 12.1, the omitted details are not cosmetic: if the analytic tensor product is strictly larger than the growth-condition product, the formula for the stalk of W^• used throughout Lemma 12.1 is wrong and the acyclicity conclusion is unsupported. Please provide a complete proof, or a precise reference, for Lemma 7.14, including the module-structure and growth estimates needed to reduce to (7.6).","section":"§7.2, Lemma 7.14"},{"comment":"The reduction of Theorem 1.10 to Lemma 12.1 is valid only if the local description supplied by Remark 7.15 and Lemma 7.14 is correct, and only if stalk-wise acyclicity at the origin in each ETD local model implies global acyclicity of K^•. The latter point is stated in one sentence ('Lemma 12.1 below shows that K^• is acyclic for all ETDs with one-dimensional base, so φ^• is a quasi-isomorphism'). Please spell out the sheaf-theoretic argument: one needs to know that the cohomology sheaves of K^• are coherent and supported on the singular locus, whose closure contains the origin in each affine toric local model, so that vanishing at the origin forces global vanishing. This is likely true, but it should be stated explicitly because the entire relative degeneration theorem rests on this step.","section":"§12.1, proof of Theorem 1.10 and Lemma 12.1"}],"minor_comments":[{"comment":"In the display η(LS_V) = (T^1_V)^×, the notation is confusing: the manuscript then refers to '(T^1_X)^×⊂T^1_X'. Please define the subsheaf of generating sections consistently and use the same subscript throughout.","section":"§5, Theorem 5.5"},{"comment":"The sentence 'We infer the notion of strata to the normalization of X' is awkward; consider rephrasing as 'We carry the notion of strata over to the normalization of X.'","section":"§6, Definition 6.6"},{"comment":"The letter h is used for a local homomorphism P→N in Lemma 7.13 and then reused in the growth conditions in (7.6) and Lemma 7.14. Please state explicitly that one fixes such an h once and for all, or explain why the growth condition is independent of the choice of h.","section":"§7.2, equation (7.6) and Lemma 7.14"},{"comment":"The proof concludes 'so φ^• is a quasi-isomorphism and Theorem 1.10 follows by the discussion in §2.1.' Since the exact sequence defining K^• involves analytic sheaves, a brief comment on the passage from local acyclicity to a global quasi-isomorphism would improve readability; see the corresponding major comment.","section":"§12.1, end of proof of Theorem 1.10"},{"comment":"The abstract mentions Frobenius manifold structures on moduli spaces as a potential application, but the body only sketches this connection. Consider softening the abstract to match the actual scope of the paper.","section":"Abstract and Introduction"}],"recommendation":"major_revision","confidential_remarks":"The paper is serious and the main theorem would be a significant advance if the proof is completed. The blocking issue is the unproved Lemma 7.14 and its role in Lemma 12.1 and Theorem 1.10; the other delegated steps are either standard or explicitly imported from published work. I recommend asking the authors to supply a complete proof of Lemma 7.14, or to replace it by a precise reference, and to expand the short gluing argument in §12.1. If those points are resolved, I would expect to support acceptance."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This is a serious paper. The main theorems are big: Danilov's Hodge–de Rham degeneration for toroidal pairs, a smoothing theorem for toroidal crossing spaces that generalizes Friedman's d-semistability and bypasses Gross–Siebert local rigidity, and a relative degeneration theorem for one-parameter log toroidal families. If the central arguments hold, this is a major advance.\n\nWhat is genuinely new: the log toroidal family framework, the explicit base-change results in positive characteristic, the Cartier isomorphism in this setting, and the use of BV/Maurer–Cartan theory to turn degeneration into smoothings. The authors also deserve credit for identifying concrete flaws in earlier work (Kawamata–Namikawa p.404, Gross–Siebert Theorem 4.1) and showing how their results repair them. The main theorems are stated cleanly, and the proof strategy is coherent.\n\nThe stress-test flags a real soft spot: Lemma 12.1, the acyclicity statement at the origin that underlies Theorem 1.10, rests on Lemma 7.14, whose proof is literally left to the reader. I read Lemma 7.14 and it does look plausible—you can reduce to finitely many subspaces and then to the scalar growth-condition estimate of (7.6)—but it is load-bearing, not a throwaway. A referee should insist the details be written out. Two milder gaps: Lemma 10.4 (Cartier isomorphism) is quite terse, and Theorem 6.13 is imported from Gross–Siebert without proof. The unproved conjecture for arbitrary coherent bases is not a real concern for the main smoothing result, since Theorem 1.7 only needs the sharp toric monoid bases of the form Spec(N -> k) and Sm.\n\nWho this is for: algebraic geometers working in mirror symmetry, log geometry, Hodge theory, or deformation theory. It deserves a serious referee—someone who will actually check Lemma 7.14 and the local toric computations. My own verdict is conditional: I lean toward the results being correct, but the current write-up leaves one too many technical details to the reader at a critical junction.\n\nRecommendation: send to peer review, with the referee explicitly asked to focus on Section 7.2 and Lemma 12.1. The authors should either prove Lemma 7.14 in full or reorganize so that its proof is clearly supplied. Once that is done, I would expect this to become an important and accepted paper.","headline":"A substantial and likely-correct paper that settles Danilov's conjecture and gives a general smoothing criterion; the main risk is a load-bearing analytic lemma whose proof is deferred to the reader.","tokens_in":40986,"tokens_out":2408,"would_cite":true,"duration_ms":23032,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14D15","14J32","14M25","14F40","14B07"],"pacs":[],"model":"deepseek-v4-flash","headline":"A toroidal crossing space with a simple log section and a transverse anticanonical divisor is smoothable to an orbifold with terminal singularities.","keywords":["toroidal crossing spaces","log structures","Hodge–de Rham degeneration","smoothing","normal crossing spaces","Calabi-Yau varieties","Maurer-Cartan equation","Batalin-Vilkovisky operator"],"falsifier":"Find a proper log toroidal family over $\\mathrm{Spec}(Q\\to k)$ with $Q$ a sharp toric monoid for which the spectral sequence $E^{p,q}_1=R^qf_*W^p_{X/S}$ does not degenerate at $E_1$, or a proper toroidal pair $(X,D)$ with $\\sum_{p+q=n}\\dim H^q(X,\\tilde\\Omega^p_X(\\log D)) > \\dim H^n(X,\\tilde\\Omega^{\\bullet}_X(\\log D))$; either would contradict the paper's central degeneration theorem.","tokens_in":39779,"feed_emoji":"📐","tokens_out":10846,"duration_ms":100230,"temperature":0.7,"pith_summary":"The paper proves that a proper toroidal crossing space can be smoothed to an orbifold with terminal singularities, provided it carries a simple section of the sheaf of log smooth structures on a dense open set and an anticanonical section whose zero divisor meets all strata transversely. The proof works by turning the log structure into an infinitesimal deformation of the singular locus, then solving a Maurer–Cartan equation in the Gerstenhaber algebra of log polyvector fields to build the formal smoothing, and finally passing to an analytic smoothing. Along the way the paper establishes the degeneration of the Hodge–de Rham spectral sequence for proper log toroidal families, settling a long-standing conjecture for toroidal pairs that the analogous spectral sequence degenerates at $E_1$. If the central theorem is correct, it produces smoothings in many situations where earlier d-semistability or local-rigidity assumptions fail, and it opens a route to constructing new Calabi–Yau and Fano varieties and Frobenius manifold structures on moduli spaces.","feed_headline":"Toroidal crossing spaces smooth out under mild assumptions","feed_subtitle":"A new degeneration proof settles a long-standing Hodge conjecture and opens Calabi–Yau construction.","key_machinery":"The central objects are the sheaf $LS_X$ of log smooth structures on a toroidal crossing space, whose sections map to the sheaf $T^1_X$ of first-order deformations, and the elementary log toroidal local models $(Q\\subset P,F)$: an injection of sharp toric monoids with $P$ a free $Q$-set and a distinguished set of facets, describing $X$ \\'etale locally near the log singular locus. On these models the paper computes the sheaves $W^p_{X/S}=j_*\\Omega^p_{U/S}$ of Zariski–Steenbrink–Danilov differentials explicitly, proves base change for them in sufficiently large characteristic, constructs the Cartier isomorphism, and obtains a Frobenius decomposition of $F_*W^{\\bullet}_{X_0/S_0}$. The degeneration theorem then follows by the spreading-out-to-finite-characteristic method. For the smoothing itself, the machinery is the Maurer–Cartan equation in the Gerstenhaber algebra of log polyvector fields, with the Batalin–Vilkovisky operator $\\Delta$ transporting the de Rham differential via a chosen volume form; Theorem 1.10 makes the relevant cohomology free, and the deformation-obstruction theorem for log toroidal families controls the lifting steps.","core_discovery":"The central claim is Theorem 1.7: a proper toroidal crossing space $X$ with a simple section $s$ of $LS_X$ on a dense open set, together with an anticanonical section whose zero divisor $E$ meets all strata of $X$ and $Z$ transversely, is smoothable to an orbifold with terminal singularities. The engine is the degeneration theorem for the Hodge–de Rham spectral sequence of a proper log toroidal family $f:X\\to S$ with $S=\\mathrm{Spec}(Q\\to k)$, where $Q$ is a sharp toric monoid: the spectral sequence $E^{p,q}_1=R^qf_*W^p_{X/S}$ converges to $R^{p+q}f_*W^{\\bullet}_{X/S}$ and degenerates at $E_1$. This degeneration implies the conjecture for proper toroidal pairs stated as Theorem 1.4, and in the relative one-parameter case over $S_m=\\mathrm{Spec}(\\mathbb{N}\\to\\mathbb{C}[t]/(t^{m+1}))$ it yields Theorem 1.10: the higher direct images are free, commute with base change, and the spectral sequence degenerates. For normal crossing spaces, the general theorem specializes to the cleaner statement that $X$ is smoothable whenever $T^1_X$ is generated by global sections and $X_{\\mathrm{sing}}$ is projective.","pith_inferences":["If the degeneration theorem extends to arbitrary coherent log bases as the paper conjectures, the same smoothing conclusion should hold for families whose base log structure is not a single sharp monoid, and toroidal-pair degeneration would cover more general boundary divisors.","The base-change failure in low characteristic (illustrated in the paper for a specific monoid in characteristic two) suggests that modular or characteristic-$p$ interpretations of these smoothings may require excluding finitely many primes; this is a concrete condition to verify in applications.","The Maurer–Cartan and Batalin–Vilkovisky construction ties the smoothing directly to a chosen anticanonical volume form, so the resulting orbifold smoothing should carry a natural log Calabi–Yau structure; testing this on the hyperplane-union example might yield an explicit Frobenius manifold structure near the boundary of the moduli space."],"forward_implications":["Normal crossing spaces with effective anti-canonical class and $T^1_X$ generated by global sections are smoothable whenever $X_{\\mathrm{sing}}$ is projective, and without projectivity a reduced section with transverse zero locus suffices.","The degeneration theorem settles the conjecture for proper toroidal pairs: the Hodge–de Rham spectral sequence for $\\tilde\\Omega^{\\bullet}_X(\\log D)$ degenerates at $E_1$.","For one-parameter log toroidal families over $S_m=\\mathrm{Spec}(\\mathbb{N}\\to\\mathbb{C}[t]/(t^{m+1}))$, the Hodge bundles $R^qf_*W^p_{X/S}$ are free and commute with base change, so Hodge bundles extend trivially over toroidal boundary divisors in moduli.","The smoothing theorem applies where earlier criteria fail, for example to unions of $d$ hyperplanes in $\\mathbb{P}^n$ with $d\\le n+1$ and to double, triple, or higher intersections of Fano components along divisors, yielding new Calabi–Yau and Fano manifolds."],"supporting_citations":[{"why":"Defines the sheaf $LS_X$ classifying log smooth structures on a toroidal crossing space and gives its embedding into normal-bundle torsors.","marker":"[21]"},{"why":"Supplies the local models along the zero locus of a simple section and the deformation-obstruction theorem used to control log deformations.","marker":"[22]"},{"why":"Provides the spreading-out, Cartier isomorphism, and Frobenius decomposition technique for proving Hodge–de Rham degeneration.","marker":"[15]"},{"why":"Gives the Maurer–Cartan and Batalin–Vilkovisky machinery that converts degeneration into an unobstructed formal deformation.","marker":"[8]"},{"why":"Provides the Cartier isomorphism for log smooth morphisms in positive characteristic, a key ingredient of the degeneration proof.","marker":"[34]"},{"why":"Supplies the relative one-parameter degeneration argument used to obtain freeness and base change for the Hodge bundles.","marker":"[48]"},{"why":"Introduces the $T^1_X$ criterion and d-semistability for normal crossing smoothability that the main theorem generalizes.","marker":"[19]"},{"why":"Gives the bridge from formal smoothings to analytic smoothings via Hilbert schemes and approximation arguments.","marker":"[46]"}],"fun_headline_variants":["Toroidal crossing spaces smooth out via log structures","Smoothing theorem settles Danilov's Hodge conjecture","New proof smooths toroidal crossing spaces","Smoothing toroidal spaces: Danilov conjecture proven","Log structures yield smoothings for crossing spaces"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument requires the family to be covered by explicit monomial local models of log toroidal type over a base whose log structure is a single sharp toric monoid, or by the one-parameter ring $\\mathbb{C}[t]/(t^{m+1})$; for more general coherent log bases the key degeneration theorem is only conjectured, so the smoothing proof does not apply there.","fun_headline_variants_meta":{"raw":{"variants":["Toroidal crossing spaces smooth out via log structures","Smoothing theorem settles Danilov's Hodge conjecture","New proof smooths toroidal crossing spaces","Smoothing toroidal spaces: Danilov conjecture proven","Log structures yield smoothings for crossing spaces"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000236,"raw_usage":{"total_tokens":1504,"prompt_tokens":947,"completion_tokens":557,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":563,"completion_tokens_details":{"reasoning_tokens":482}},"tokens_in":563,"tokens_out":557,"duration_ms":5421,"temperature":1.0,"reasoning_tokens":482,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T10:21:31.508700+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Find a proper log toroidal family over $\\mathrm{Spec}(Q\\to k)$ with $Q$ a sharp toric monoid for which the spectral sequence $E^{p,q}_1=R^qf_*W^p_{X/S}$ does not degenerate at $E_1$, or a proper toroidal pair $(X,D)$ with $\\sum_{p+q=n}\\dim H^q(X,\\tilde\\Omega^p_X(\\log D)) > \\dim H^n(X,\\tilde\\Omega^{\\bullet}_X(\\log D))$; either would contradict the paper's central degeneration theorem.","supporting_citations":[{"cited_title":"Mirror symmetry via logarithmic degeneration data","cited_arxiv_id":null,"evidence_quote":"Defines the sheaf $LS_X$ classifying log smooth structures on a toroidal crossing space and gives its embedding into normal-bundle torsors."},{"cited_title":"Mirror symmetry via logarithmic degeneration data, II","cited_arxiv_id":null,"evidence_quote":"Supplies the local models along the zero locus of a simple section and the deformation-obstruction theorem used to control log deformations."},{"cited_title":"Rel` evements modulo p2 et d´ ecomposition du complexe de de Rham","cited_arxiv_id":null,"evidence_quote":"Provides the spreading-out, Cartier isomorphism, and Frobenius decomposition technique for proving Hodge–de Rham degeneration."},{"cited_title":"Geometry of the Maurer-Cartan equation near degenerate Calabi-Yau varieties","cited_arxiv_id":"1902.11174","evidence_quote":"Gives the Maurer–Cartan and Batalin–Vilkovisky machinery that converts degeneration into an unobstructed formal deformation."},{"cited_title":"Logarithmic structures of Fontaine-Illusie","cited_arxiv_id":null,"evidence_quote":"Provides the Cartier isomorphism for log smooth morphisms in positive characteristic, a key ingredient of the degeneration proof."},{"cited_title":"Steenbrink","cited_arxiv_id":null,"evidence_quote":"Supplies the relative one-parameter degeneration argument used to obtain freeness and base change for the Hodge bundles."},{"cited_title":"Global smoothings of varieties with normal crossings","cited_arxiv_id":null,"evidence_quote":"Introduces the $T^1_X$ criterion and d-semistability for normal crossing smoothability that the main theorem generalizes."},{"cited_title":"Period integrals from wall structures via tropical cycles, canonical coordinates in mirror symmetry and analyticity of toric degenerations","cited_arxiv_id":null,"evidence_quote":"Gives the bridge from formal smoothings to analytic smoothings via Hilbert schemes and approximation arguments."}],"review_version":1}