{"id":"b2d99f25-9a38-4a18-8a21-d33d12895b3e","arxiv_id":"2412.11583","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Every complex analytic singularity admitting a contracting automorphism is quasi-homogeneous.","lead":"This mathematics paper proves that any complex singularity admitting a contracting self-map must be quasi-homogeneous, meaning it has a scaling symmetry. The result extends a previously known two-dimensional characterization to all dimensions.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Proof of Theorem 3.3 assumes existence of a holomorphic curve in X through p whose loop lies in the contracting basin, without proof.","rationale":"The reader's weakest_assumption identifies exactly the load-bearing gap I find: the unproved existence of a holomorphic curve in X through p whose loop is inside a contracting neighborhood. The paper's Theorem 3.3 depends on this inference to prove that any extension F of a contracting automorphism is contracting on the ambient space. Without it, the later use of Poincaré-Dulac normal forms and the invariant-ideal argument in Theorem 4.1 do not get started. I also considered the secondary issue of passing from λ-homogeneous to standard weighted homogeneous generators; this is a real missing detail but is far less serious because a positive integer weight vector orthogonal to the relation lattice always exists when the relation lattice contains no nonzero nonnegative vector, which follows from |λ_i|<1. The curve-selection issue is more delicate: for arbitrary possibly reducible singularities, no curve-selection lemma is stated, and the contraction property for arbitrary compact sets is not a consequence of Definition 2.1. The proof is likely repairable, but as written it is conditional; the reader's verdict of CONDITIONAL with MODERATE confidence is appropriate. My stress-test does not change that verdict.","tokens_in":14117,"tokens_out":27256,"duration_ms":261670,"concrete_test":"Write out and prove a curve-selection lemma: for a reduced analytic germ (X,0) not contained in a proper analytic subgerm S, for every neighborhood U of 0 there exist p∈(X\\S)∩U and a holomorphic map φ:(C,0)→(X,0) with φ(0)=0, p∈φ(Δ), and φ(Δ)⊂U. Check the lemma on the reducible example X={xy=0}⊂C^2 with S={x=0} and p=(1,0): does every such p admit a curve inside X with image in an arbitrarily small U? If the lemma can be proved for arbitrary reduced germs, the gap in Theorem 3.3 is repairable by inserting it; if a counterexample exists, the proof of the theorem is invalid.","verdict_should_be":"UNCHANGED","load_bearing_attack":"In §3.2, after choosing p∈ι(X,0) outside the stable manifold, the paper states: \"Then we can find a holomorphic curve φ: (C,0)→(C^d,0) such that φ(0)=0 and p∈Im(φ).\" It does not assert that Im(φ)⊂ι(X,0), nor that C=φ(S^1) lies in a neighborhood on which f acts contractively. The subsequent step \"Let C=φ(S^1). As it is compact and F is contracting on X, we know that for all ε>0 there exists m such that F^m(C)⊂B(0,ε)\" requires that forward iterates of the loop C converge to 0 in X. Definition 2.1 only provides an attracting neighborhood U with F(U)⋐U and ∩F^n(U)=0; it does not state that every compact set in X is swallowed. One must both select a curve lying in X and choose p close enough to 0 so that C lies inside such a U. Neither is proved. If this curve-selection and basin-membership step fails, the estimate |c_l λ_l^m| ≤ ε/2π no longer follows, and the contradiction |λ_l|<1 collapses. This is the only place where the contracting hypothesis on X is used in Theorem 3.3, so the theorem is not fully established as written.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper claims Theorem A: if a complex analytic singularity (X,0) admits an automorphism f that is contracting in the sense of Definition 2.1, then (X,0) is quasi-homogeneous. The proof has two main steps. First, Theorem 3.3 asserts that in an embedding of minimal dimension (a \"good\" embedding), any extension of f to the ambient (C^d,0) is automatically a contracting automorphism of the ambient space. This is proved by contradiction using the stable manifold theorem and a holomorphic curve through a point outside the stable manifold. Second, Theorem 4.1 asserts that any ideal invariant under a contracting automorphism of (C^d,0) is generated by weighted homogeneous polynomials. This is proved by introducing a λ-order and λ-gradation adapted to the linear part of the automorphism, and then running a double induction on the order and the number of generators. The paper concludes that the ideal of (X,0) is weighted homogeneous, hence the singularity is quasi-homogeneous.","tokens_in":14337,"tokens_out":44483,"duration_ms":401339,"significance":"If correct, this is a strong result: it extends the two-dimensional results of Favre–Ruggiero and Camacho–Movasati–Scárdua to arbitrary dimensions and to arbitrary (possibly reducible) singularities, giving a complete dynamical characterization of quasi-homogeneity. The λ-gradation machinery in Section 4 is a potentially useful combinatorial tool for invariant ideals under contracting automorphisms. The paper relies on standard external results (stable manifold theorem, Poincaré–Dulac normal forms), has no free parameters, and does not appear to be circular. However, several load-bearing steps in the proofs are not justified as written, so the significance is conditional on these gaps being repaired.","major_comments":[{"comment":"The proof chooses a holomorphic curve φ: (C,0)→(C^d,0) with φ(0)=0 and p∈Im(φ), but it does not show that Im(φ)⊂ι(X,0). The subsequent step \"Let C=φ(S^1). As it is compact and F is contracting on X, we know that for all ε>0 there exists m such that F^m(C)⊂B(0,ε)\" applies the contraction property of F on X to the loop C, which is only justified if C⊂ι(X,0) and C lies inside an attracting neighborhood U as in Definition 2.1. Since p is chosen outside the stable manifold, it is not automatic that a curve in X through 0 and p exists or that its trace on a circle around 0 is contained in the basin of attraction. Please provide a proof of the required curve selection inside X (e.g., via desingularization or normalization) and explain how p is chosen small enough so that C⊂U. Without this, the estimate leading to |λ_l|<1 does not follow.","section":"Theorem 3.3, §3.2"},{"comment":"The proof that every contracting automorphism is uniformly contracting is defective. The sequence (x_n) is not shown to converge; only a subsequence converges, and the limit of a sequence in X\\V need not lie outside V (it may lie on the boundary). The statement itself is true, and a correct proof can be obtained by noting that the compact sets F^n(K) are nested with intersection {0}, so they must eventually be contained in any neighborhood of 0. Since the uniform contraction property is later used to justify that the compact set C is swallowed by arbitrarily small balls, this proof should be rewritten.","section":"§2.1, Proposition 2.3"},{"comment":"The claim that the λ-order makes (Γ,⪰) order isomorphic to (N,≥) is false for generic eigenvalues. When the numbers |λ_i| are rationally independent (e.g., λ_1=2^{-√2}, λ_2=2^{-√3}), the set {|λ^{-α}|: α∈N^d} is dense in [1,∞), so the λ-order is not a well-order. Consequently, the induction over Γ in Proposition 4.22, which refers to \"the preceding pair (j,δ)\", is not justified; a transfinite induction would be needed, and the order type is not N. This is a load-bearing issue for the combinatorial proof of Theorem 4.1, since the invariant-ideal argument depends on this induction.","section":"§4.2, Definition 4.7 and Remark 4.10(2)"},{"comment":"Theorem 4.1 concludes that I is generated by weighted homogeneous polynomials in the sense of Definition 2.6, which requires a single integer weight vector (n_1,...,n_d) with n_i≠0. The proof, however, establishes generation by λ-homogeneous polynomials with respect to the gradation H^γ. It is not shown that these polynomials are weighted homogeneous for a common integer weight vector. A necessary additional step is to prove that the lattice L={m∈Z^d: λ^m=1} has a rational orthogonal complement containing a vector with all coordinates nonzero; this follows from |λ_i|<1, but it is not mentioned. Please add this argument, as Definition 2.7 and Theorem A depend on it.","section":"Theorem 4.1 and §4.5–4.6"}],"minor_comments":[{"comment":"The Cauchy estimate has a numerical factor error: the final bound should be ε, not ε/(2π). This does not affect the conclusion that |λ_l|<1, but it should be corrected.","section":"§3.2, proof of Theorem 3.3"},{"comment":"The text says \"two n-tuples α, α′\" but the ambient dimension is d; this should read \"d-tuples\".","section":"§4.2, Definition 4.8"},{"comment":"The auxiliary map used to define the partial order is denoted φ, which conflicts with the holomorphic curve φ in Theorem 3.3. Using a different letter would avoid confusion.","section":"§4.4, Lemma 4.26"},{"comment":"The proof of Proposition 4.22 uses ϕ^(i)(0)=0, which is only justified if I is an ideal of functions vanishing at 0, i.e., the ideal of a subvariety through 0. If Theorem 4.1 is intended for arbitrary invariant ideals, this restriction should be stated explicitly; otherwise, the theorem statement should say I⊂m_{C^d,0}.","section":"Theorem 4.1 and §4.5"},{"comment":"The manuscript does not explicitly state that the singularity (X,0) is reduced. Several arguments (e.g., analytic continuation of the stable manifold intersection, use of open subsets of X) implicitly require reducedness. Please state the reducedness assumption or adjust the proofs to cover the non-reduced case.","section":"Throughout"},{"comment":"There is a typo in the heading of Section 4: \"Inv ariant ideals\" should be \"Invariant ideals\".","section":"Section title"}],"recommendation":"major_revision","confidential_remarks":"The paper addresses a natural and important question, and the overall strategy is attractive. However, the current manuscript contains multiple load-bearing gaps: the curve-selection and basin-membership step in Theorem 3.3, the flawed proof of Proposition 2.3, the non-well-founded λ-order used in the induction for Theorem 4.1, and the missing passage from λ-homogeneous to classical weighted homogeneous generators. The first two may be repairable by simpler arguments (notably, a direct stable-manifold/tangent-space argument seems to avoid the curve entirely), but the λ-order issue is more structural and requires reworking the induction in Section 4. I recommend major revision and encourage the author to address these points carefully."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Kémo Morvan has a real theorem here: it removes the dimension restriction from the Favre-Ruggiero result and shows that a contracting automorphism forces quasi-homogeneity in arbitrary dimension. That is a natural open question and this is a genuine advance. The proof strategy is sensible—extend to a minimal embedding, use Poincaré-Dulac normal form, then analyze invariant ideals via a λ-gradation. Section 4 is the real novelty: the gradation by eigenvalue products and the double-inclusion argument for ideals of arbitrary codimension are thoughtful and mostly convincing.\n\nThe two soft spots the reader flagged are real but smaller than they look. In Theorem 3.3, the curve φ is not explicitly required to lie inside X until the moment it is used. That is a one-line fix: take the attracting neighborhood U from Definition 2.1, pick p in U ∩ ι(X,0), and choose a curve through p inside U. The contraction property then gives the swallowed loop exactly as claimed. The paper does not say this, and a referee should ask for it, but it is not a load-bearing error.\n\nThe second worry, about passing from λ-homogeneous generators to standard weighted homogeneous generators, also has a clean patch. The equivalence classes are the cosets of the resonance lattice L = {α−β : λ^α = λ^β}. Since |λ_i|<1, L contains no nonzero vector in the nonnegative orthant, so L^⊥ contains a vector with all positive integer entries. That vector gives the weight system. The paper does not spell this out; it just asserts the conclusion. Same diagnosis: fixable, needs a few lines.\n\nI do not see circularity or hidden fitting. The citations to Favre-Ruggiero, Orlik-Wagreich, and the normal form literature are appropriate. The main theorem is the genuine route to the result.\n\nThis paper deserves a serious referee. It is a strong contribution to singularity theory and holomorphic dynamics. The referee will need to ask for the two clarifications above, and maybe for a careful rewriting of Proposition 2.3, but the core is in good shape. I would send it out.","headline":"A genuine higher-dimensional generalization of Favre-Ruggiero; the main proof is sound, with two fixable gaps in exposition.","tokens_in":14868,"tokens_out":11302,"would_cite":true,"duration_ms":95486,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["32S05","37F99","32B10","14B05","32A10"],"pacs":[],"model":"deepseek-v4-flash","headline":"Any complex singularity with a contracting automorphism is quasi-homogeneous.","keywords":["contracting automorphism","quasi-homogeneous singularity","weighted homogeneous polynomial","holomorphic dynamics","C*-action","Poincaré-Dulac normal form","invariant ideal","complex analytic singularity"],"falsifier":"An explicit counterexample would be a non-quasi-homogeneous singularity, for instance a non-reduced germ or a union of branches with incompatible weights, that still admits a contracting automorphism; the first place to look is where the proof's curve-selection step fails, namely a reducible germ in which no holomorphic arc through a point off the shrinking direction lies entirely in the singularity.","tokens_in":13886,"feed_emoji":"🌀","tokens_out":8654,"duration_ms":67815,"temperature":0.7,"pith_summary":"This paper proves that a complex singularity that admits a contracting automorphism—a holomorphic self-map fixing the origin, invertible, and shrinking a whole neighborhood into itself so that every orbit converges to the origin—must be quasi-homogeneous, meaning it can be embedded so that its defining ideal is generated by weighted homogeneous polynomials. This concludes, in all dimensions and without normality assumptions, a characterization that was previously known only for normal surface singularities and for two-dimensional Stein spaces. The result connects the dynamical existence of a contracting symmetry to the existence of a $C^*$-action on the singularity. The proof avoids resolution of singularities: it embeds the germ minimally, extends the automorphism to a contracting automorphism of the ambient space, and then shows that any invariant ideal under such an ambient contraction is generated by weighted homogeneous polynomials.","feed_headline":"Contracting automorphisms force quasi-homogeneous singularities","feed_subtitle":"In any dimension, a germ with such a self-map must be defined by weighted homogeneous equations.","key_machinery":"The argument is carried by two devices. The first is the theory of good embeddings: when a singularity germ is embedded in $\\mathbb{C}^d$ with $d$ equal to its embedding dimension, every extension of an automorphism of the germ is an ambient automorphism, and every extension of a contracting automorphism is an ambient contracting automorphism (Theorem 3.3). The proof of the contracting part uses the stable manifold theorem together with Cauchy estimates on holomorphic arcs that leave the stable manifold, forcing all eigenvalues of the ambient derivative to have modulus $<1$. The second device is the $\\lambda$-gradation built from the spectrum of a contracting ambient automorphism in Poincaré-Dulac normal form: the $\\lambda$-order on $\\mathbb{N}^d$ organizes monomials into finite-dimensional spaces $H^\\gamma$ of $\\lambda$-homogeneous polynomials, and lemmas 4.15 and 4.16 show that these spaces are permuted and stably acted on by the automorphism. The main algebraic result (Theorem 4.1) uses a double inclusion and a triangular linear algebra lemma to prove that any invariant ideal is generated by such $\\lambda$-homogeneous polynomials.","core_discovery":"The central claim is Theorem A: if $(X,0)$ is a complex analytic singularity and $f:(X,0)\\to(X,0)$ is a contracting automorphism, then $(X,0)$ is quasi-homogeneous. Since every quasi-homogeneous singularity carries contracting automorphisms through its $C^*$-action, the theorem gives a full characterization of when contracting automorphisms exist. The proof splits into two statements that are of independent interest. Theorem 3.3 shows that in an embedding of minimal dimension, every extension of a contracting automorphism of the singularity is a contracting automorphism of the ambient space. Theorem 4.1 shows that any ideal of holomorphic functions invariant under a contracting diffeomorphism of $\\mathbb{C}^d$ in Poincaré-Dulac normal form is generated by weighted homogeneous polynomials, with an invariant filtration of the ideal preserved. Applied to the defining ideal of the singularity, these two statements yield Theorem A.","pith_inferences":["A natural testable extension is to replace 'automorphism' by 'endomorphism': the Poincaré-Dulac and invariant-ideal arguments appear to need only a contracting self-map, so the same conclusion may hold without invertibility.","The $\\lambda$-gradation machinery could be applied to ideals invariant under a semigroup of commuting contractions, which would give a new way to prove quasi-homogeneity for singularities with a $\\mathbb{C}^*$-action without resolving them.","Combining Theorem 4.1 with Saito's criterion would make every singularity with a contracting automorphism have equal Milnor and Tjurina numbers, and conversely a dynamical proof of quasi-homogeneity from that equality.","Since the proof avoids resolution, it may carry over to formal or real-analytic germs where resolution is unavailable, provided the stable manifold theorem and Poincaré-Dulac normal form have analogues."],"forward_implications":["Every complex singularity that admits a contracting automorphism is quasi-homogeneous, so it also carries a $C^*$-action and weighted-homogeneous defining equations.","The dimension-two results of Favre and Ruggiero and of Camacho, Movasati, and Scárdua become special cases of a single theorem valid for arbitrary germs.","Theorem 4.1 provides a normal form for invariant ideals under contracting diffeomorphisms: any such ideal has a filtered generating set of $\\lambda$-homogeneous polynomials.","Theorem 3.3 shows that the embedding dimension is the correct invariant for extension problems: in minimal embeddings, contractions of the germ extend to contractions of the ambient space."],"supporting_citations":[{"why":"Establishes the dimension-2 case for normal surface singularities that this paper generalizes to arbitrary germs.","marker":"[6]"},{"why":"Classifies quasi-homogeneous singularities with a $C^*$-action and supplies the embedding strategy used here.","marker":"[10]"},{"why":"Treats the two-dimensional Stein case and shows which surface singularities support the relevant dynamics.","marker":"[4]"},{"why":"Provides the stable manifold theorem used in the proof of Theorem 3.3.","marker":"[1]"},{"why":"Gives the Poincaré-Dulac normal form for contracting diffeomorphisms used in Section 4.","marker":"[13]"},{"why":"Supplies the embedding dimension criterion and the result that minimal embeddings are good for automorphisms.","marker":"[7]"},{"why":"States the Milnor–Tjurina criterion for quasi-homogeneity, motivating the comparison in the introduction.","marker":"[12]"}],"fun_headline_variants":["Contracting automorphisms characterize quasi-homogeneous singularities","Contracting self-maps force weighted homogeneous singularities","A singularity with a contracting automorphism is quasi-homogeneous","Contracting automorphisms: a full quasi-homogeneity criterion","Quasi-homogeneous iff it admits a contracting automorphism"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that for every point of the singularity off the shrinking direction one can draw a holomorphic arc inside the singularity from the origin through that point, and the automorphism shrinks the whole circle trace of that arc uniformly; this curve selection is assumed rather than proven for arbitrary (possibly reducible) singularities.","fun_headline_variants_meta":{"raw":{"variants":["Contracting automorphisms characterize quasi-homogeneous singularities","Contracting self-maps force weighted homogeneous singularities","A singularity with a contracting automorphism is quasi-homogeneous","Contracting automorphisms: a full quasi-homogeneity criterion","Quasi-homogeneous iff it admits a contracting automorphism"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001064,"raw_usage":{"total_tokens":4366,"prompt_tokens":756,"completion_tokens":3610,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":372,"completion_tokens_details":{"reasoning_tokens":3525}},"tokens_in":372,"tokens_out":3610,"duration_ms":24587,"temperature":1.0,"reasoning_tokens":3525,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T14:50:58.789526+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"An explicit counterexample would be a non-quasi-homogeneous singularity, for instance a non-reduced germ or a union of branches with incompatible weights, that still admits a contracting automorphism; the first place to look is where the proof's curve-selection step fails, namely a reducible germ in which no holomorphic arc through a point off the shrinking direction lies entirely in the singularity.","supporting_citations":[{"cited_title":"Normal surface singularities admitting contracting automorphisms.Ann","cited_arxiv_id":null,"evidence_quote":"Establishes the dimension-2 case for normal surface singularities that this paper generalizes to arbitrary germs."},{"cited_title":"Isolated singularities of algebraic surfaces withC∗ action","cited_arxiv_id":null,"evidence_quote":"Classifies quasi-homogeneous singularities with a $C^*$-action and supplies the embedding strategy used here."},{"cited_title":"Movasati, and Bruno Scardua","cited_arxiv_id":null,"evidence_quote":"Treats the two-dimensional Stein case and shows which surface singularities support the relevant dynamics."},{"cited_title":"An introduction to hyperbolic dynamical systems","cited_arxiv_id":null,"evidence_quote":"Provides the stable manifold theorem used in the proof of Theorem 3.3."},{"cited_title":"Local contractions and a theorem of poincaré","cited_arxiv_id":null,"evidence_quote":"Gives the Poincaré-Dulac normal form for contracting diffeomorphisms used in Section 4."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the embedding dimension criterion and the result that minimal embeddings are good for automorphisms."},{"cited_title":"Quasihomogene isolierte singularitäten von hyperflächen.Inventiones mathematicae, 14:123–142, 1971","cited_arxiv_id":null,"evidence_quote":"States the Milnor–Tjurina criterion for quasi-homogeneity, motivating the comparison in the introduction."}],"review_version":1}