{"id":"a566e5ea-1aa5-4942-bbb8-a2de0ef2eaf0","arxiv_id":"2411.15896","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Trace, norm, and the new central divisor fully characterize slice regular functions over H and R_3 up to holomorphic conjugation by the automorphism group.","lead":"This paper proves that two slice regular functions over the quaternions, or over the Clifford algebra R_3, are equivalent under a holomorphic family of algebra automorphisms exactly when they share three invariants: trace, norm, and a newly introduced central divisor. A generalist reader may care because this closes the natural equivalence problem in quaternionic analysis and sharpens an earlier classification that required working with a larger class of functions.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The proof of (ii)⇒(iv) in §18 never verifies that the common factor λ produced by Proposition 15.1 is conjugation-symmetric, so F̃ and H̃ need not be stem functions before Theorem 18.1 is applied; this missing lemma is the main repair needed.","rationale":"The reader's CONDITIONAL verdict is appropriate. The central theorem is plausible and the proof strategy — local section, isotropy vector field, sheaf-cohomology globalization — is sound in outline. The most load-bearing concrete problem I found is not the isotropy-group structure that the reader emphasized as the weakest assumption; that part of the argument appears correct, since all isotropy groups of SO(3, C) on W⊗C\\{0} are connected and one-dimensional, and the sheaf S is indeed locally isomorphic to Z, allowing Proposition 16.2 to give H^1(D, A) = 0. Instead, the key step (ii)⇒(iv) in §18 has a missing conjugation-symmetry check: Proposition 15.1 supplies an arbitrary holomorphic λ, but Theorem 18.1 requires stem functions. The fix is standard — choose λ symmetric under conjugation by pairing conjugate divisor factors — but it is neither stated nor proved. This is a repairable gap, not a fatal flaw, so the verdict should remain CONDITIONAL. The R3 statement also has a small definitional imprecision concerning cdiv for slice-preserving components, but the underlying mathematics is consistent once Tr and N are seen to force such components to agree, so I did not elevate this to the main concern.","tokens_in":23428,"tokens_out":45118,"duration_ms":435587,"concrete_test":"Re-derive the factorization step in §18 with the additional requirement λ(̄z) = ̄λ(z). Since cdiv(F) = cdiv(H) is a symmetric divisor on the symmetric domain D, construct λ explicitly: for each non-real zero pair {a, ̄a} of multiplicity m take (z−a)^m(z−̄a)^m, for each real zero r of multiplicity m take (z−r)^m, and use symmetric Weierstrass products if the divisor is infinite. Then verify that F̃ = Ĥ/λ and H̃ = Ĥ/λ satisfy the stem-function reality condition and avoid zero, so Theorem 18.1 applies. If this construction always works on conjugation-symmetric domains in C, insert it as a lemma before §18; if a domain is found with a symmetric divisor admitting no conjugation-symmetric generator, the proof of (ii)⇒(iv) needs a different argument.","verdict_should_be":"UNCHANGED","load_bearing_attack":"In §18, after reducing to Ĥ = (Ĥ − Ĥ^c)/2 and Ĥ = (H − H^c)/2, the proof uses cdiv(F) = cdiv(H) and Proposition 15.1 to obtain holomorphic functions Ĥ = λ F̃ and Ĥ = λ H̃ with F̃, H̃ : D → A_C \\ {0}. Proposition 15.1 only guarantees a holomorphic scalar λ whose divisor is the common divisor; it does not guarantee the stem-function reality condition λ(̄z) = ̄λ(z). Without that condition, F̃ and H̃ need not be stem functions, and Theorem 18.1 — which explicitly assumes stem functions — cannot be invoked. This is a genuine gap in the central implication (ii)⇒(iv). It is repairable: because cdiv(F) is the divisor of the W-component of a stem function on a symmetric domain, its multiplicities at z and ̄z agree, so a conjugation-symmetric generator λ can be built by pairing conjugate zero factors (z−a)^m(z−̄a)^m and using Weierstrass products for infinite divisors. But this argument is absent from the manuscript. The other globalization steps — the one-dimensional connected isotropy groups, the triviality of the isotropy Lie-algebra bundle, and the H^1 vanishing via Proposition 16.2 — appear coherent, and the R3 mixed slice-preserving boundary case is also consistent once one notes that a non-slice-preserving quaternionic stem function with the same Tr and N as a slice-preserving one is forced to be slice-preserving by the reality condition on real points.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the action of the automorphism group of the quaternion algebra H (and of the Clifford algebra R3 ≅ H⊕H) on slice regular functions defined on axially symmetric domains. The main result (Theorem 1.1) states that for two non-slice-preserving slice regular functions f,h with stem functions F,H, the following are equivalent: the invariants (cdiv, Tr, N) coincide; the stem functions have the same invariants; for each z the values F(z) and H(z) lie in the same Aut(HC)-orbit; there is a holomorphic map φ:D→Aut(HC) with F=φ(H); and there is a holomorphic α:D→H_C^* with F=α^{-1}Hα. An analogous statement is given for R3 (Theorem 8.1). The proof combines the stem-function formalism of Ghiloni–Perotti, the orbit structure of SO(3,C) on the imaginary part of HC, local-to-global holomorphic section theorems (Grauert, Forstnerič, Matsushima–Morimoto), and a cohomology vanishing result for a locally constant Z-sheaf.","tokens_in":23652,"tokens_out":19565,"duration_ms":164641,"significance":"If correct, the paper settles the automorphism classification for slice regular functions over H and R3 in a clean way: trace, norm, and the newly introduced central divisor form a complete set of invariants, and the equivalence with holomorphic conjugacy is stronger than the previously known semi-regular results of Altavilla–de Fabritiis. The paper is technically ambitious, combining stem functions with Stein theory and relative symmetric products; the explicit examples in §3 and §4 usefully demonstrate why the cdiv invariant is needed. The proof of Theorem 1.1 relies on standard advanced tools (Grauert's existence results, Oka-principle type arguments), and the overall structure is broadly coherent. The main theorem for H appears defensible modulo the repair described below.","major_comments":[{"comment":"After constructing λ, F̃, H̃ via Proposition 15.1, the paper applies Theorem 18.1, which explicitly assumes that the maps are stem functions (F(z̄)=conj(F(z)) and H(z̄)=conj(H(z))). The functions F̃ and H̃ are only known to be holomorphic maps into AC\\{0} with Tr=0 and equal norms; no argument is given that λ can be chosen conjugation-symmetric, so the stem-function hypothesis is not verified. This is a genuine gap in the central implication. It is repairable in two ways: either prove that the common divisor can be generated by a conjugation-symmetric λ (pairing zeros at z and z̄), or observe that the proof of Theorem 18.1 nowhere uses the stem-function condition and restate Theorem 18.1 without that hypothesis. As written, the invocation of Theorem 18.1 is unjustified.","section":"§18 (proof of Theorem 1.1(a), (ii)⇒(iv))"},{"comment":"The central divisor for R3 is defined componentwise, but in §6.2 cdiv is defined only for stem functions that are not slice-preserving. Theorem 8.1 states an equivalence for arbitrary f,h:Ω_D→R3, yet condition (3) is not well-formed when a component is slice-preserving. The one-line proof does not discuss the mixed case (one component slice-preserving, the other not), nor how Theorem 1.1(b) would supply the missing implication. The theorem should either explicitly restrict to functions whose components are all non-slice-preserving, or it should state and prove the slice-preserving component case separately.","section":"Theorem 8.1 and §6.4"}],"minor_comments":[{"comment":"There is a missing reference placeholder '[?]' in the list of related references in the introduction; the sentence should not be left with an unresolved citation.","section":"Introduction"},{"comment":"The first line of the proof is garbled ('1 = 1 · ¯1 = ⇒ 1 = ...'); it should be rewritten to show correctly that the antiinvolution fixes 1.","section":"§2.1, proof of Lemma 2.2"},{"comment":"The proposition is stated for 'short exact sequence of OX module sheaves', but the sheaf S is only a locally constant Z-sheaf, not an OX-module; the cohomological argument uses abelian sheaf cohomology, so the statement should be rephrased accordingly.","section":"§16, Proposition 16.2"},{"comment":"The proof of Proposition 13.1 is very terse: it asserts without detail that the isotropy Lie algebras form a holomorphic line bundle and that a nowhere-vanishing section yields a flow with the required properties. A few more sentences explaining the construction would improve readability and verifiability.","section":"§13, Proposition 13.1"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is a substantial contribution, but the two gaps identified above—the unverified stem-function hypothesis in §18 and the missing hypothesis/undefined cdiv in Theorem 8.1 for mixed slice-preserving components—need to be fixed before publication. The first gap is easily repaired either by strengthening Theorem 18.1 or by proving conjugation symmetry of λ; the second requires a careful restatement for R3. The presence of a citation placeholder suggests the manuscript is not yet in final form, and a careful revision of the proof of Theorem 8.1 is advisable."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe short version: the paper is a real contribution, but the key implication in §18 has a repairable gap. The authors introduce a new invariant, the central divisor, and prove that trace, norm, and cdiv completely classify slice regular functions over H and R3 up to holomorphic conjugation. That is genuinely new and sharpens the semi-regular classification from [ADF20b]. The examples in §3–§4 are worked out and show cdiv is not redundant.\n\nThe problem is in the proof of (ii)⇒(iv). After cutting to imaginary parts, the authors use Proposition 15.1 to write ˆF = λ F̃ and ˆH = λ H̃. Proposition 15.1 only gives a holomorphic λ with the right divisor; it does not ensure λ(z̄) = \\overline{λ(z)}. Without that, F̃ and H̃ are not stem functions, so Theorem 18.1 does not apply. This is a real gap in the written proof. It is repairable: the common divisor is symmetric because it comes from stem functions, so one can construct a conjugation-symmetric λ by pairing conjugate zeros (or using Weierstrass products for infinite divisors). But the argument is not there.\n\nEverything else I checked looks coherent. The complex-geometry machinery — relative symmetric products, the tubular neighborhood retraction, the sheaf cohomology with isotropy sheaves — is used correctly as far as I can tell, and the reliance on external results keeps the circularity burden low. The other weaknesses are minor: two '[?]' placeholders, four uncited references, and a terse norm identity in the same section.\n\nVerdict: conditional, leaning positive. The missing lemma is standard to fix, not a structural failure. This paper is for specialists in slice regular functions and Stein geometry; it deserves a serious referee. My recommendation: send it to review, ask the authors to add the conjugation-symmetry argument and clean up the references.","headline":"A genuinely new classification theorem for slice regular functions over H and R3, with a repairable gap in the proof of the key implication (ii)⇒(iv).","tokens_in":24347,"tokens_out":9284,"would_cite":true,"duration_ms":75008,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["30G35","15A66","16W20"],"pacs":[],"model":"deepseek-v4-flash","headline":"Trace, norm, and a central divisor form a complete set of automorphism invariants for slice regular functions over the quaternions and the Clifford algebra R3.","keywords":["slice regular functions","quaternions","Clifford algebra R3","central divisor","stem functions","trace","norm","automorphism invariants"],"falsifier":"Find two non-slice-preserving slice regular functions with equal trace, norm, and central divisor, yet no holomorphic map $\\varphi:D\\to\\operatorname{Aut}(\\mathbb H_{\\mathbb C})$ with $F(z)=\\varphi(z)(H(z))$; the theorem asserts none exists. A concrete route is to compute the 1-cocycle obtained from local holomorphic solutions on a two-open-set cover of $D$ and check whether it is a coboundary in the sheaf $\\mathcal A$ of holomorphic functions into the isotropy groups. For the explicit pair in Section 4—$F(z)=I+zJ+\\frac12 z^2K$ and $G(z)=(1+\\frac12 z^2)I$, which have equal trace and norm but different central divisor—the theorem predicts that no holomorphic $\\alpha:D\\to\\mathbb H_{\\mathbb C}^*$ satisfies $F=\\alpha^{-1}G\\alpha$; a direct power-series solution would settle that prediction.","tokens_in":23083,"feed_emoji":"🔄","tokens_out":13907,"duration_ms":113523,"temperature":0.7,"pith_summary":"The paper proves that two slice regular functions on a symmetric domain over the quaternions, or over the Clifford algebra $\\mathbb{R}_3\\cong\\mathbb{H}\\oplus\\mathbb{H}$, are equivalent under a holomorphically varying automorphism exactly when they agree in three invariants: pointwise trace, pointwise norm, and a newly introduced \"central divisor.\" The central divisor $\\operatorname{cdiv}(F)$ records where the stem function $F:D\\to\\mathbb{H}_{\\mathbb C}$ takes values in the center of the complexified algebra, and it is defined precisely when the function is not slice preserving. This upgrades the classical pointwise statement for quaternions—where trace and norm determine the automorphism orbit of a single value—to a statement about functions, where the automorphism must be allowed to vary holomorphically over the domain. The extra divisor is needed because pointwise automorphisms can always be chosen, but they can be glued into one global holomorphic family only away from the zeros of the non-central part of the stem function. If the theorem is right, matching these three invariants is both necessary and sufficient for two slice regular functions to lie in the same orbit of the natural automorphism action.","feed_headline":"Three invariants decide slice regular function equivalence","feed_subtitle":"Equal trace, norm, and central divisor guarantee a holomorphic conjugation.","key_machinery":"The machinery has three parts. First, stem functions: every slice regular function on $\\Omega_D$ is encoded by a holomorphic function $F:D\\to\\mathbb H_{\\mathbb C}$ satisfying $F(\\bar z)=\\overline{F(z)}$, and automorphisms of the function algebra act on these stems, with trace and norm passing through the correspondence exactly. Second, the central divisor: writing $\\mathbb H_{\\mathbb C}=\\mathbb C\\oplus(W\\otimes\\mathbb C)$ with $W$ the imaginary subspace of $\\mathbb H$, the stem function splits as $F=(F',F'')$, and $\\operatorname{cdiv}(F)$ is the divisor of $F'':D\\to W\\otimes\\mathbb C$, recording where $F$ enters the center. Third, the gluing mechanism: for $F,H$ with the same invariants, one forms $V=\\{(z,g)\\in D\\times G_{\\mathbb C}:F(z)=g(H(z))\\}$ with $G_{\\mathbb C}=\\operatorname{Aut}(\\mathbb H_{\\mathbb C})$. The fibers of $V\\to D$ are the isotropy groups of the action on the imaginary directions; they are all one-dimensional, and generically isomorphic to $\\mathbb C^*$. Because the generic isotropy group is commutative and one-dimensional, the relevant isotropy sheaf is locally isomorphic to $\\mathbb Z$, and a sheaf-cohomology vanishing result ($H^1(D,\\mathcal A)=0$ for the sheaf $\\mathcal A$ of holomorphic sections into isotropy groups) converts local holomorphic sections into a global one. The same structure, applied componentwise, covers $\\mathbb{R}_3$.","core_discovery":"The central result is Theorem 1.1. Let $D\\subset\\mathbb C$ be a symmetric domain, $\\Omega_D$ the corresponding axially symmetric domain in the quaternions, and let $f,h:\\Omega_D\\to\\mathbb H$ be slice regular functions with stem functions $F,H:D\\to\\mathbb H_{\\mathbb C}$. If neither $f$ nor $h$ is slice preserving—meaning neither stem function collapses into the center of $\\mathbb H_{\\mathbb C}$—then the following are equivalent: (i) $f$ and $h$ have the same trace, norm, and central divisor; (ii) $F$ and $H$ have the same trace, norm, and central divisor; (iii) $F(z)$ and $H(z)$ lie in the same $\\operatorname{Aut}(\\mathbb H_{\\mathbb C})$-orbit for every $z\\in D$; (iv) there is a holomorphic map $\\varphi:D\\to\\operatorname{Aut}(\\mathbb H_{\\mathbb C})$ with $F(z)=\\varphi(z)(H(z))$ for all $z$; (v) there is a holomorphic map $\\alpha:D\\to\\mathbb H_{\\mathbb C}^*$ with $F(z)=\\alpha(z)^{-1}H(z)\\alpha(z)$. The passage from (iii) to (iv) is the real content: pointwise equivalence of values is promoted to global holomorphic conjugacy, and (v) shows that this conjugacy is inner, matching the Skolem–Noether picture for the constant algebra. If $f$ is slice preserving, the statement collapses to $f=h$, since automorphisms fix the center pointwise. Theorem 8.1 proves the analogous equivalence for slice regular functions with values in $\\mathbb{R}_3\\cong\\mathbb{H}\\oplus\\mathbb{H}$, using the connected component $\\operatorname{Aut}(\\mathbb H_{\\mathbb C})\\times\\operatorname{Aut}(\\mathbb H_{\\mathbb C})$ of the automorphism group.","pith_inferences":["A consequence the paper leaves implicit is that the method is tailored to algebras whose isotropic stabilizers are connected and one-dimensional; extending the classification to other Clifford algebras will likely require new invariants or a different gluing argument, and the authors announce the investigation in a forthcoming paper.","The central divisor behaves like an analogue of the classical divisor of a holomorphic function: it records the zeros of the non-central part of the stem function, and the theorem says that this divisor together with trace and norm is the complete obstruction to holomorphic conjugacy. This suggests defining a relative central divisor for functions that are slice preserving only on a subset, and te","Since the proof shows the automorphism action on the function algebra is inner, one could ask whether the invariant description transfers to functions on domains with boundary or to other real alternative algebras with connected one-dimensional isotropy groups; the analytical ingredients are largely independent of the specific algebra.","The introduction's link to orthogonal complex structures could be tested concretely: if slice regular functions in the same automorphism orbit induce the same or isomorphic orthogonal complex structures, then the three invariants would give a practical way to detect equivalence of such structures."],"forward_implications":["Two non-slice-preserving quaternionic slice regular functions are holomorphically conjugate by an inner automorphism-valued map exactly when their trace, norm, and central divisor agree; this gives a complete, computable invariant for the automorphism action.","The same classification holds for $\\mathbb{R}_3$-valued slice regular functions with respect to the connected automorphism group; for the full automorphism group, trace and norm are invariants only up to the order-reversing swap of the two $\\mathbb{H}$ factors.","Slice-preserving functions are rigid: if one of the two functions is slice preserving, the equivalence collapses to equality of the functions themselves, because automorphisms fix the center.","The result strengthens the semi-regular equivalence theory: conjugation can be achieved by a genuinely slice regular invertible function rather than only by a semi-regular one, at the cost of adding the central divisor condition.","For individual values in $\\mathbb H_{\\mathbb C}$, the theorem recovers the classical fact that equality of trace and norm is equivalent to lying on the same $\\operatorname{Aut}(\\mathbb H_{\\mathbb C})$-orbit."],"supporting_citations":[{"why":"Supplies the stem-function correspondence for slice regular functions over real alternative algebras, the bridge between functions and their complexified stems.","marker":"[GP11]"},{"why":"Proves the semi-regular equivalence via trace, norm, and the Sylvester operator; it is the baseline that this paper strengthens to holomorphic conjugacy by regular functions.","marker":"[ADF20b]"},{"why":"Source for the classical description of $\\operatorname{Aut}(\\mathbb H)\\cong SO(3,\\mathbb R)$ acting on the imaginary subspace, used throughout the paper.","marker":"[CS03]"},{"why":"Provides the reduced trace and norm and the Skolem–Noether background used to identify automorphisms of $\\mathbb H$ and $\\mathbb H_{\\mathbb C}$ with inner conjugations.","marker":"[Lor08]"},{"why":"Provides the holomorphic tubular-neighborhood and retraction theorem used in Proposition 9.4 to obtain local holomorphic sections of the fiber bundle $V\\to D$.","marker":"[For17]"},{"why":"Establishes that simply connected complex Lie groups are Stein manifolds, used to justify the local-section construction in Proposition 9.4.","marker":"[MM60]"}],"fun_headline_variants":["Three invariants decide slice regular function conjugation","Same trace, norm, divisor: slice regular functions are conjugate","Holomorphic conjugacy for slice regular functions: triple test","Three numbers tell if slice regular functions are conjugate"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that every non-real direction in the complexified algebra is fixed by a connected one-dimensional group of automorphisms—generically the nonzero complex numbers—so that local choices of automorphism can be glued into one global holomorphic choice.","fun_headline_variants_meta":{"raw":{"variants":["Three invariants decide slice regular function conjugation","Same trace, norm, divisor: slice regular functions are conjugate","Holomorphic conjugacy for slice regular functions: triple test","Three numbers tell if slice regular functions are conjugate"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000999,"raw_usage":{"total_tokens":4245,"prompt_tokens":977,"completion_tokens":3268,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":593,"completion_tokens_details":{"reasoning_tokens":3214}},"tokens_in":593,"tokens_out":3268,"duration_ms":19434,"temperature":1.0,"reasoning_tokens":3214,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T13:50:41.951564+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Find two non-slice-preserving slice regular functions with equal trace, norm, and central divisor, yet no holomorphic map $\\varphi:D\\to\\operatorname{Aut}(\\mathbb H_{\\mathbb C})$ with $F(z)=\\varphi(z)(H(z))$; the theorem asserts none exists. A concrete route is to compute the 1-cocycle obtained from local holomorphic solutions on a two-open-set cover of $D$ and check whether it is a coboundary in the sheaf $\\mathcal A$ of holomorphic functions into the isotropy groups. For the explicit pair in Section 4—$F(z)=I+zJ+\\frac12 z^2K$ and $G(z)=(1+\\frac12 z^2)I$, which have equal trace and norm but different central divisor—the theorem predicts that no holomorphic $\\alpha:D\\to\\mathbb H_{\\mathbb C}^*$ satisfies $F=\\alpha^{-1}G\\alpha$; a direct power-series solution would settle that prediction.","supporting_citations":[],"review_version":1}