{"id":"f18c1b1a-c16d-4416-92fd-b28988097ddc","arxiv_id":"2411.17143","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"SAut_k(A^2) is simple as an infinite-dimensional algebraic group over any infinite field k; closed normal subgroups of SAut_k(A^n) contain all tame automorphisms.","lead":"This paper proves that the group of polynomial automorphisms of the affine plane with Jacobian 1 is topologically simple: over any infinite field, its only closed normal subgroups are the trivial group and the whole group. It also shows that in every dimension, any closed normal subgroup contains all tame automorphisms, giving a new path toward the higher-dimensional problem.","discovery_kind":"first_principles","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Proposition 4.2 is false when g has only constant x-terms; affine non-translation automorphisms are not covered by Proposition 4.3, leaving a repairable gap in the proof of Proposition B.","rationale":"I read the paper in good faith and believe the main theorem is true: the overall strategy is elegant, and the non-affine case is handled convincingly using the topological specialization property. However, the written proof contains a false auxiliary statement: Proposition 4.2 breaks down when g has only constant terms in x, which is exactly what happens when f is affine but not a translation. The reader's weakest assumption concerned the topological specialization property, which is a different point; my concern is a concrete internal gap in Proposition 4.2/4.3. This gap does not falsify the central claim, because the affine case can be patched with a one-line commutator argument, but it means the manuscript as submitted does not fully prove Proposition B as claimed. I therefore recommend a conditional acceptance: the authors should add the missing affine case or amend Proposition 4.2 to allow a = 0 and explicitly treat translations over k[t].","tokens_in":17541,"tokens_out":12380,"duration_ms":115639,"concrete_test":"Run the construction of Proposition 4.3 for f = (x1+x2, x2) in SAut_k(A^2) and τ = (x1, x2+t). This yields g = (x1-t, x2), whose expansion has only m = 0 terms; check that no coprime a,b ≥ 1 satisfy a/b = max{m/j} = 0 and that h_ε(0) = id for all ε, contradicting Proposition 4.2(2). Then verify that g(1) = (x1-1, x2) is a nontrivial translation lying in the normal subgroup generated by f, which confirms the intended theorem is unchanged once the affine case is added explicitly.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"In Proposition 4.2, the proof writes g_i = x_i + sum_{j≥1,m≥0} q_{i,j,m} t^j with q_{i,j,m} homogeneous of degree m, then chooses coprime integers a,b ≥ 1 with a/b = max{m/j | q_{i,j,m} ≠ 0}. If every nonzero q_{i,j,m} has m = 0, this maximum is 0, so no such a,b ≥ 1 exist. This case actually occurs in the proof of Proposition 4.3: when f is affine but not a translation, τ^{-1}f^{-1}τ f is a k[t]-translation. For example, with f = (x1+x2, x2) and τ = (x1, x2+t), one gets g = (x1-t, x2). For every a,b ≥ 1 and every ε, h_ε(0) = id, so Proposition 4.2(2) fails. Thus Proposition 4.3, stated for every f not in Tran(k), has no valid proof for affine non-translation f, and Proposition B has a gap for such f. The gap is benign: for affine f, any translation σ not commuting with f gives a nontrivial translation σ^{-1}f^{-1}σ f in N, so the proof can be repaired by handling the affine case separately. But the manuscript does not do this, so the proof of Theorem A is incomplete as written.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proves Theorem A: for an infinite field k and n≥1, every closed normal subgroup N of SAut_k(A^n) is either trivial or contains all tame automorphisms of SAut_k(A^n); in particular SAut_k(A^2) is simple as an ind-group. The proof strategy is to show (Proposition B) that a nontrivial closed normal subgroup contains all translations, and then to invoke and reproduce Lewis's theorem (Proposition C) that translations generate all tame automorphisms. The translation step is the new ingredient: using the one-parameter family g = τ^{-1} f^{-1} τ f with τ a k[t]-translation, a limit argument produces a non-trivial translation in the closure of N. The paper also treats finite fields (Proposition D) and gives a result for Aut_k(A^n) (Proposition E).","tokens_in":17700,"tokens_out":9990,"duration_ms":82605,"significance":"If the proof is completed, Theorem A resolves a long-standing open problem on the simplicity of SAut_k(A^2) as an ind-group and provides the first general statement for higher-dimensional special automorphism groups. The paper is largely self-contained: it reproduces the relevant results of Lewis, gives explicit computations in Sections 3 and 4, and cleanly isolates the topological specialization property needed for the limit argument. The finite-field results and the Aut_k(A^n) proposition are valuable complements. However, the proof of Proposition B currently has a gap (see Major Comment 1), so the central claim is not yet established as written.","major_comments":[{"comment":"Proposition 4.2 is false as stated. If all non-zero q_{i,j,m} have m=0, the maximum of m/j is 0, so no coprime integers a,b ≥ 1 exist with a/b = max{m/j}; moreover, for any a,b ≥ 1, the specialization h_ε(0) is the identity for every ε. The example g = (x_1 - t, x_2, ..., x_n) illustrates this: h_ε = (x_1 - t^a, x_2, ..., x_n), so h_ε(0) = id. This case occurs in Proposition 4.3 when f is affine but not a translation, since then g = τ^{-1} f^{-1} τ f is a k[t]-translation; for f = (x_1+x_2, x_2) and τ = (x_1, x_2+t), g = (x_1 - t, x_2). Consequently the proof of Proposition 4.3, and hence of Proposition B, does not cover affine non-translation f. The gap is repairable: if f is affine and does not commute with a translation σ, then σ^{-1} f^{-1} σ f is a non-trivial translation in N. The manuscript should either prove Proposition 4.2 under a hypothesis that excludes this case (and then handle the affine case separately) or adjust the statement.","section":"§4.1, Proposition 4.2"}],"minor_comments":[{"comment":"The word 'inclustion' should be 'inclusion'.","section":"§2"},{"comment":"The phrase 'There exists moreover a a surjective group homomorphism' contains a duplicated article 'a'.","section":"Introduction, Proposition D"},{"comment":"In the definition of h_{q,ǫ}, the expression 'q(x2, . . . , qn)' should read 'q(x2, . . . , xn)'.","section":"§3.3, proof of Proposition 3.6"},{"comment":"In the characteristic-2 case for n=2, the translation is written as '(x1 + θµν(θ^2 + µ^2), x2, )' with a stray comma; please clarify the intended composition.","section":"§3.3, proof of Proposition 3.6"},{"comment":"The notation 'g(ta)' should be 'g(t^a)' with a superscript, to avoid confusion with the product 't a'.","section":"§4.1, Proposition 4.2"},{"comment":"In the induction step, 'the homogeneous part of degree i of ∑...' should be 'degree j', not 'degree i'.","section":"§4.2, Lemma 4.1"},{"comment":"The topological specialization property (if h(t0) ∈ N for all t0 ∈ k*, then h(0) ∈ N for a closed normal N) is used essentially in Proposition B but is asserted rather than proved or referenced for the two ind-topologies; please add a proof or a precise reference showing that k* is dense in the relevant topology on A^1.","section":"Introduction and §4.1, Proposition B"}],"recommendation":"major_revision","confidential_remarks":"The gap in Proposition 4.2 is real but appears straightforwardly repairable by treating affine non-translation automorphisms separately in Proposition 4.3. The rest of the manuscript is careful and the main theorem, if completed, would be a significant result. I therefore recommend major revision rather than rejection; the author should also be asked to provide the missing justification for the topological specialization property used in Proposition B."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Main take: this is a serious paper with a real result, but the proof as written has a genuine gap that needs patching. The affine non-translation case in Proposition 4.3 is not covered by Proposition 4.2.\n\nWhat's new: Theorem A resolves a long-open question. The earlier proof by Shafarevich used a false Lie correspondence, and Furter–Kraft's result gives closed normal subgroups of Aut(A^n) rather than SAut. The paper's one-parameter limit construction (Props 4.2–4.3) and the closedness argument (Prop B) are new and mostly clean. Reproving Lewis's generation results makes the paper self-contained; the finite-field section is a nice bonus with concrete quotient maps to (k[x],+). No free parameters, no circularity, and the citation pattern looks appropriate.\n\nThe soft spot: Proposition 4.2 chooses a,b with a/b = max(m/j). If all nonzero q_{i,j,m} have m=0, no such coprime positive integers exist. That case is not exotic—it happens exactly for affine non-translation f in Prop 4.3: e.g. f=(x1+x2,x2), tau=(x1,x2+t) gives g=(x1-t,x2). For such g, h_epsilon(0)=id for all epsilon, so Prop 4.2(2) fails and Prop 4.3 has no valid proof for affine f. The gap is repairable: if f is affine non-translation, take a translation sigma not commuting with f; then sigma^{-1} f^{-1} sigma f is a nontrivial translation in N. But the manuscript doesn't do this, so the proof of Prop B is incomplete as written. The topological specialization property is also stated rather than proved, and some ind-topology facts are used without proof; those are standard and likely fine.\n\nBottom line: the main theorem is probably correct, and the fix is small. For people in affine algebraic geometry and ind-groups, this is worth reading and deserves a serious referee. I would send it to review and expect acceptance after minor revision.","headline":"The main result is likely true and the strategy is mostly sound, but Proposition 4.2 has a genuine unhandled case (affine non-translations) that needs a small repair.","tokens_in":18413,"tokens_out":4204,"would_cite":true,"duration_ms":36014,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14R10","14R20"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper proves that SAut_k(A^2) is simple as an ind-group over any infinite field, and that closed normal subgroups in higher dimensions contain all tame automorphisms.","keywords":["automorphism groups of affine space","ind-groups","simple group","tame automorphisms","closed normal subgroups","polynomial automorphisms","Nagata automorphism","finite fields"],"falsifier":"Find an infinite field k and a proper nontrivial closed normal subgroup of SAut_k($A^{2}$). Equivalently, find an element f of SAut_{k[t]}(A^n) and a closed normal subgroup N such that the specialized family h(t_0) belongs to N for every t_0 ∈ k^* but h(0) does not; such an example would show the specialization property fails and would block Theorem A.","tokens_in":17184,"feed_emoji":"🔁","tokens_out":5288,"duration_ms":43475,"temperature":0.7,"pith_summary":"The paper proves that the group SAut_k($A^{2}$) of polynomial automorphisms of the affine plane with Jacobian 1 is simple as an infinite-dimensional algebraic group, for every infinite field k: its only closed normal subgroups are the trivial group and the whole group. In higher dimensions, the same argument shows that every nontrivial closed normal subgroup of SAut_k(A^n) contains all tame automorphisms. This resolves a question that had been open since a 1981 proof by Shafarevich was found to rely on a false Lie correspondence. The paper also shows the behavior is completely different over finite fields, where SAut_k($A^{2}$) admits surjective homomorphisms onto the additive group of k[x].","feed_headline":"The automorphism group of the affine plane is topologically simple","feed_subtitle":"A proof that every closed normal subgroup of Jacobian-one plane automorphisms is trivial or the whole group.","key_machinery":"The load-bearing construction is a family of commutators h = $ρ^{{-1}}$ g(t^a) ρ built from a non-translation f and suitable translations, where replacing t by a nonzero element of k gives elements of the form ρ($τ^{{-1}}$ $f^{{-1}}$ τ f)$ρ^{{-1}}$ that lie in the closed normal subgroup N by normality. Because the map t ↦ h(t) is continuous for the ind-group topology on SAut_k(A^n), the preimage of N is closed; it contains k^*, so by the specialization property it contains 0, and h(0) is a nontrivial translation. This forces N to contain all translations, and Proposition C (due to Lewis) then forces N to contain SL_n(k) and all tame automorphisms. The ind-group structure, in which automorphisms of bounded degree form finite-dimensional affine varieties, is what makes the specialization property available.","core_discovery":"Blanc establishes Theorem A: for any infinite field k and any integer n ≥ 1, if N is a closed normal subgroup of SAut_k(A^n), then either N = {id} or N contains every tame automorphism. For n = 2, since every automorphism is tame, this makes SAut_k($A^{2}$) simple as an ind-group. The proof first shows that any nontrivial closed normal subgroup contains a nontrivial translation, then uses a result of Lewis to lift translations to SL_n(k) and then to all tame automorphisms. The finite-field case is treated separately, yielding surjective homomorphisms to the additive group of k[x] whose kernels contain all translations.","pith_inferences":["The limited role of the topology suggests that the same specialization axiom could be a useful test for simplicity of other infinite-dimensional automorphism groups, such as ind-subgroups of the Cremona group.","Because the proof shows that the closure of commutator conjugates of any non-translation contains a translation, similar rigidity may hold for closed normal subgroups of other groups of polynomial automorphisms, for instance those preserving a symplectic form.","Over finite fields, the non-perfectness of SAut_k(A^2) comes from a quotient onto k[x]/V; a natural question is whether the kernels of these quotients are closed in an ind-topology adapted to finite fields, though the usual topology degenerates to discrete.","The method likely extends to automorphism groups of A^n over rings or semi-local bases, as long as the specialization property and the centralizer lemma hold."],"forward_implications":["Any nontrivial closed normal subgroup of SAut_k(A^2) is the whole group, over any infinite field k.","For n ≥ 3, every nontrivial closed normal subgroup of SAut_k(A^n) contains all tame automorphisms; in particular, by Example 3.8, it contains the Nagata automorphism.","If N is a closed normal subgroup of Aut_k(A^n) with Jacobian image equal to k^*, then N = Aut_k(A^n) (Proposition E).","Over finite fields, SAut_k(A^2) is not perfect and is not finitely generated, and the normal subgroup generated by affine automorphisms is proper (Proposition D).","The argument isolates a purely topological condition, so the same conclusion holds for any topology on SAut_k(A^n) satisfying the specialization property, such as the two standard ind-topologies or, for local fields, the topology induced by the field."],"supporting_citations":[{"why":"Supplies Proposition C: any normal subgroup of SAut(A^n) containing all translations contains SL_n(k), and containing SL_n(k) contains all tame automorphisms; this is the propagation step.","marker":"[Lew20]"},{"why":"Provides the ind-group structure and topology on automorphism groups, and earlier results on closed normal subgroups of Aut(A^n) that the paper extends.","marker":"[FK18]"},{"why":"Earlier result that conjugacy classes of most elements of SAut(A^2) are closed; the paper's proof of Proposition B builds on this circle of ideas.","marker":"[Bla16]"},{"why":"Shows SAut(A^2) is not simple as an abstract group, highlighting that closedness is the key hypothesis.","marker":"[Dan74]"},{"why":"Introduces the Nagata automorphism used in Example 3.8 to show that nontrivial closed normal subgroups in dimension three contain non-tame elements.","marker":"[Nag72]"},{"why":"Provides the explicit conjugation computation showing the Nagata automorphism lies in the normal subgroup generated by SL_3(k).","marker":"[MP09]"},{"why":"Shows existence of non-tame automorphisms in three variables over characteristic zero, motivating the higher-dimensional statement.","marker":"[SU04]"}],"fun_headline_variants":["Plane automorphism group proven topologically simple","No nontrivial closed normal subgroups in affine plane automorphisms","Simple as an ind-group: automorphisms of the affine plane","All closed normal subgroups of plane automorphisms are trivial or whole","Affine plane automorphisms: only trivial closed normal subgroups"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof rests on the topological specialization property: if a closed normal subgroup contains h(t) for every nonzero t in k, then it contains the limit h(0) obtained by setting t = 0. If a topology on SAut_k(A^n) fails this property, the nontrivial translation that drives the whole argument could fall outside the subgroup.","fun_headline_variants_meta":{"raw":{"variants":["Plane automorphism group proven topologically simple","No nontrivial closed normal subgroups in affine plane automorphisms","Simple as an ind-group: automorphisms of the affine plane","All closed normal subgroups of plane automorphisms are trivial or whole","Affine plane automorphisms: only trivial closed normal subgroups"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000701,"raw_usage":{"total_tokens":3063,"prompt_tokens":741,"completion_tokens":2322,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":357,"completion_tokens_details":{"reasoning_tokens":2236}},"tokens_in":357,"tokens_out":2322,"duration_ms":15577,"temperature":1.0,"reasoning_tokens":2236,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T12:29:29.924782+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Find an infinite field k and a proper nontrivial closed normal subgroup of SAut_k($A^{2}$). Equivalently, find an element f of SAut_{k[t]}(A^n) and a closed normal subgroup N such that the specialized family h(t_0) belongs to N for every t_0 ∈ k^* but h(0) does not; such an example would show the specialization property fails and would block Theorem A.","supporting_citations":[],"review_version":1}