{"id":"4f340c4b-dcd2-40ed-9a98-aed3ecdd5f55","arxiv_id":"2608.04940","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Every symmetric fusion category over R is equivalent to the semi-linear super representation category of a Z2-graded finite super group.","lead":"The paper proves a complete classification of finite real symmetry categories, the real-number analogue of Deligne's classification over the complex numbers. It also gives a dictionary between these categories and certain finite groupoid data.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Main theorem hinges on unproved classification of semi-linear actions: Proposition 3.15 (and hence Lemma 6.3) is not established, so the reduction to group extensions is not fully justified.","rationale":"The paper's main theorem is important and probably true; the overall strategy, complexify, apply Deligne's theorem, classify semi-linear Z2-actions, and descend, is coherent and mostly uses standard tools. I agree with the reader that the weakest point is the classification of semi-linear Z2-actions on sRep_C(K,z). The proof of this classification hinges on Proposition 3.15 and Lemma 6.3, which are not fully proved. I did not find a specific counterexample, and the bosonic analogue (Proposition 3.14) is standard, but the omitted proof is not cosmetic: it is exactly the claim that every symmetric monoidal autoequivalence of sRep_C(K,z) comes from a BZ2-equivariant autoequivalence of BK. Without it, the step 'by Proposition 3.15, this super group extension induces a Z2-action' and the converse uniqueness of extensions are unsupported. The Ω_C≃C identification in Section 6.3 is a genuine local error, but it is repairable by taking the grading homomorphism s to be trivial, so it does not threaten the theorem. Therefore the reader's conditional verdict is appropriate; my read does not move it.","tokens_in":32453,"tokens_out":24169,"duration_ms":221142,"concrete_test":"Independently prove Proposition 3.15 in the minimal fermionic case K=Z2×Z2 with z=(1,0): compute Aut_{BZ2}(BK) and Aut^{E3}_C(sRep_C(K,z)) explicitly at all homotopy levels, and verify that the natural map is fully faithful and essentially surjective. In particular, for every symmetric monoidal autoequivalence F of sRep_C(K,z), show that the induced autoequivalence of the fiber-functor groupoid F_{(K,z)} is BZ2-equivariant and descends to an automorphism of BK fixing z. If a single F fails this descent, it gives a semi-linear Z2-action not produced by a Z2-graded super group extension, and Theorem 6.17 would need revision. Equivalently, formalize the omitted proof and check that Lemma 6.3's splitting map is a monoidal 2-functor.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The load-bearing step is Section 6.3's reduction of arbitrary semi-linear Z2-actions on sRep_C(K,z) to Z2-graded super group extensions. This reduction needs two facts: (i) Lemma 6.3: Aut^{E3}_{C/R}(sRep_C(K,z)) ≃ Aut^{E3}_C(sRep_C(K,z)) × Z2, so every anti-linear autoequivalence is isomorphic to a C-linear one composed with complex conjugation; and (ii) Proposition 3.15: Aut_{BZ2}(BK) ≃ Aut^{E3}_C(sRep_C(K,z)). Proposition 3.15 is asserted with its proof omitted ('we omit the details'), and Lemma 6.3 is declared 'completely identical' to the bosonic case, whose proof itself relies on Proposition 3.14. If either statement failed, there would exist semi-linear Z2-actions not of the form T_σ∘conj with T induced by a group extension, and Theorem 6.17 would not follow. The bosonic analogue is standard and the super case is plausible, but the omitted proof is exactly the point where arbitrary actions are shown to be group-theoretic, so the classification is conditional on it. Separately, the Ω_C≃C case in Section 6.3 contains a concrete error: it identifies sRep_C(G,z) with sRep_{C/R}((G,z)×Z2^T), changing the unit endomorphism algebra from C to R. This is repairable by taking s trivial, so it is not the main obstruction.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper claims a complete classification of symmetric fusion categories over the real numbers: every such category is equivalent to sRep_{C/R}(G,z,s) for some finite Z2-graded super group (Theorem 6.17). The proof proceeds by complexification, using a Galois-descent equivalence 2Vec_C^{Z2} ≃ 2Vec_R (Theorem 4.14), Deligne's classification over C, and a computation of semi-linear Z2-actions on sRep_C(K,z) in terms of Z2-graded group extensions. A second main result (Theorem 6.22) is a Tannaka-Krein style opposite equivalence between symmetric multi-fusion categories over R and finite groupoids with a Z2 × BZ2-action. The paper also characterizes which real symmetric fusion categories admit fiber functors to Vec_R or sVec_R.","tokens_in":32716,"tokens_out":11101,"duration_ms":101042,"significance":"If the technical gaps identified below are repaired, this is a substantial contribution. It provides the expected real analogue of Deligne's theorem and identifies all real symmetric fusion categories as representation categories of finite super groups with an anti-unitary grading, unifying examples such as Vec_R, Vec_H, and sVec_R. The Galois-descent framework is appropriate, and the argument is not circular: it imports Deligne's algebraically closed classification and reduces the real case to descent data. The explicit computation of fiber functors and the reconstruction theorem are additional strong results. The paper is also careful with separability and Morita-theoretic details, which is a genuine strength.","major_comments":[{"comment":"The displayed equivalence sRep_C(G,z) ≃ sRep_{C/R}((G,z)×Z2^T) is false. In sRep_C(G,z) the unit endomorphism algebra is C, whereas in sRep_{C/R}((G,z)×Z2^T) the anti-linear Z2^T factor acts on the unit by complex conjugation, forcing End(1) ≅ R. The Ω_C ≅ C case must instead be represented by a triple with trivial grading s, so that sRep_{C/R}(G,z,s) = sRep_C(G,z). This is repairable, but as written the proof of Theorem 6.17 contains an incorrect identification in a load-bearing case.","section":"6.3, first paragraph"},{"comment":"Proposition 3.15 is asserted with the proof omitted ('we omit the details'), yet it is load-bearing. It is used, through Lemma 6.3 and Proposition 6.13, to reduce arbitrary semi-linear Z2-actions on sRep_C(K,z) to Z2-graded super group extensions, which is exactly the step that yields Theorem 6.17. The manuscript should either provide the full parallel proof or cite a precise reference where the super-case equivalence is proved.","section":"3.3, Proposition 3.15"},{"comment":"The claim that the proof of Lemma 6.3 is 'completely identical' to Lemma 6.1 is not sufficiently justified. Unlike the bosonic case, the proof requires the omitted Proposition 3.15, and it also needs to check that the isomorphism F∘conj ≃ conj∘F can be chosen coherently for all C-linear autoequivalences F of sRep_C(K,z). Because this lemma is the bridge between semi-linear Z2-actions and group extensions, a complete proof should be spelled out.","section":"6.1, Lemma 6.3"}],"minor_comments":[{"comment":"For s trivial, the statement that 'tilde F_C is equivalent to Z2 × BG' should be clarified as a disjoint union of two copies of BG, rather than a product groupoid with two objects, to avoid confusion with the later homotopy quotient.","section":"6.4, after Proposition 6.20"},{"comment":"The category Vec_H is used without definition; please define it explicitly as the category of finite-dimensional modules over the quaternion algebra H.","section":"Examples 4.12 and 5.11"},{"comment":"The isomorphism H ⊗_C H ≃ M_2(C) is terse; spelling out the chosen complex structure on H would make the fusion rule H ⊗_C H ≃ C^{⊕4} transparent.","section":"Example 5.14"},{"comment":"The assertion that a field with finite-degree algebraic closure is either algebraically closed or real closed is the Artin-Schreier theorem; adding a reference would be helpful.","section":"Section 6.5, Conjecture 6.26"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is within the scope of math.QA. The two main concerns are repairable: the Ω_C ≅ C equivalence in Section 6.3 needs correction, and Proposition 3.15 needs a full proof or a precise citation. I do not see evidence of a deeper fatal flaw."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Worth reading: this is the real analogue of Deligne's theorem, and it looks basically right, with one local error and one proof gap worth pinning down. The main theorem—every symmetric fusion category over R is sRep_{C/R}(G,z,s) for a Z2-graded finite super group—is a genuine new result; Deligne's theorem stops at algebraically closed fields, and the descent machinery here is the natural next step. The Tannaka-Krein correspondence over R (Theorem 6.22) is also new and useful, and the paper handles the T^2=±1 distinction cleanly. The Galois descent framework in Section 4 is a real contribution: Theorem 4.14 gives a clean 2-categorical statement, and the proof via Morita equivalence with BMod_{C|C}(Vec_R) is well-conceived, even if some details are terse.\n\nThe soft spots. First, Section 6.3 contains a concrete error in the Ω_C ≃ C case: it claims sRep_C(G,z) is equivalent to sRep_{C/R}((G,z)×Z_2^T). That cannot be right as written—the unit endomorphism algebra of the first is C and of the second is R. The repair is local and does not threaten the theorem: take s to be trivial, or state the Ω_C ≃ C case separately as sRep_C(G,z). The paper should fix this before publication.\n\nSecond, the reduction of arbitrary semi-linear Z2-actions to group extensions relies on Proposition 3.15 (Aut_{BZ2}(BG) ≃ Aut^{E3}_C(sRep_C(G,z))), and the proof is omitted as \"parallel to Proposition 3.14.\" That is exactly the step that turns arbitrary autoequivalences into group-theoretic data, so omitting the details is uncomfortable. The statement is plausible and likely true—the bosonic version is standard, and the fiber-functor argument of Proposition 3.14 should adapt—but in a classification paper whose main theorem depends on it, this deserves a written proof or at least a precise reference. The stress-test note is right to flag it; it is not a manufactured objection.\n\nThe citation pattern looks clean: Deligne, Deligne-Milne, Etingof-Gelaki, EGNO, and the one self-citation [HXZ24] in Remark 6.16 is used for homotopy quotients and is not load-bearing. No fitted parameters, no circularity.\n\nWho this is for: anyone working on tensor categories over non-algebraically closed fields, and physicists who want a categorical home for time-reversal symmetry. It deserves a serious referee; with the Section 6.3 typo fixed and Proposition 3.15 spelled out, I would accept it. I'd bring it to reading group and would cite it.","headline":"A solid, genuinely new real analogue of Deligne's classification, with one local error in Section 6.3 and one load-bearing proof omission in Section 3 that both need fixing before it is fully rigorous.","tokens_in":33299,"tokens_out":2633,"would_cite":true,"duration_ms":23299,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["18M20","18M05","18M15","18M25"],"pacs":[],"model":"deepseek-v4-flash","headline":"Every symmetric fusion category over the real numbers is equivalent to a semi-linear super representation category of a finite $\\mathbb{Z}_2$-graded super group.","keywords":["symmetric fusion categories","super groups","Z2-graded groups","semi-linear representations","Galois descent","real forms","equivariantization","fiber functors"],"falsifier":"Compute, for the super group $(K,z)=(\\mathbb{Z}_2,1)$, i.e. for $\\mathrm{sVec}_{\\mathbb{C}}$, all semi-linear $\\mathbb{Z}_2$-actions and their equivariantizations; the paper's classification predicts exactly two actions, corresponding to the extensions $\\mathbb{Z}_2\\times\\mathbb{Z}_2$ and $\\mathbb{Z}_4$, and a third action or a third equivariantization would contradict Theorem 6.17.","tokens_in":32186,"feed_emoji":"","tokens_out":17655,"duration_ms":140134,"temperature":0.7,"pith_summary":"The paper establishes the real-number analogue of the classical classification of symmetric fusion categories over algebraically closed fields of characteristic zero: every symmetric fusion category over $\\mathbb{R}$ is equivalent to $\\mathrm{sRep}_{\\mathbb{C}/\\mathbb{R}}(G,z,s)$ for some finite super group $(G,z)$ equipped with a grading $s:G\\to\\mathbb{Z}_2$ separating unitary from anti-unitary elements. Such an object is the category of finite-dimensional complex super vector spaces on which even elements of $G$ act $\\mathbb{C}$-linearly and odd elements act anti-linearly, with the central element $z$ acting as fermion parity, so the result says that every finite real quantum symmetry, bosonic or fermionic, unitary or time-reversal-like, is concrete. The proof uses Galois descent over $\\mathbb{C}/\\mathbb{R}$: a real category is recovered from its complexification by equivariantization with respect to a semi-linear $\\mathbb{Z}_2$-action, and the paper classifies these actions on the complex categories $\\mathrm{sRep}_{\\mathbb{C}}(K,z)$ as $\\mathbb{Z}_2$-graded group extensions $1\\to K\\to G\\to\\mathbb{Z}_2\\to 1$. A further Tannaka-Krein type theorem identifies real symmetric multi-fusion categories with finite groupoids carrying a $\\mathbb{Z}_2\\times B\\mathbb{Z}_2$-action, and the paper determines exactly which real categories admit fiber functors to $\\mathrm{Vec}_{\\mathbb{R}}$ or $\\mathrm{sVec}_{\\mathbb{R}}$.","feed_headline":"Every real symmetric fusion category is a super group representation","feed_subtitle":"Proof routes real symmetries through complex ones; anti-unitary time-reversal symmetries now fit the same framework.","key_machinery":"The carrying mechanism is Galois descent over $\\mathbb{C}/\\mathbb{R}$, summarized as the symmetric monoidal equivalence of 2-categories $2\\mathrm{Vec}_{\\mathbb{C}}^{\\mathbb{Z}_2}\\simeq 2\\mathrm{Vec}_{\\mathbb{R}}$ (Theorem 4.14): real forms of a finite semisimple complex category are precisely semi-linear $\\mathbb{Z}_2$-actions, i.e. anti-linear symmetric monoidal autoequivalences $J$ with $J^2\\cong 1$. The classification step is the computation, for each finite super group $(K,z)$, of all such actions on $\\mathrm{sRep}_{\\mathbb{C}}(K,z)$: Section 6.2 shows that a semi-linear $\\mathbb{Z}_2$-action is the same datum as a $\\mathbb{Z}_2$-graded extension $1\\to K\\to G\\to\\mathbb{Z}_2\\to 1$ with $z$ central, and equivariantization produces $\\mathrm{sRep}_{\\mathbb{C}/\\mathbb{R}}(G,z,s)$. In the reconstruction half, the carrying object is the groupoid $\\widetilde{\\mathcal{F}}_{\\mathcal{C}}$ of $\\mathbb{R}$-linear fiber functors $\\mathcal{C}\\to\\mathrm{sVec}_{\\mathbb{C}}$, equipped with the action of the 2-group $\\mathrm{Aut}_{\\mathbb{R}}(\\mathrm{sVec}_{\\mathbb{C}})\\simeq\\mathbb{Z}_2\\times B\\mathbb{Z}_2$; Theorem 6.21 is the reconstruction $\\mathcal{C}\\simeq\\mathrm{Fun}_{\\mathbb{Z}_2\\times B\\mathbb{Z}_2}(\\widetilde{\\mathcal{F}}_{\\mathcal{C}},\\mathrm{sVec}_{\\mathbb{C}})$.","core_discovery":"The central claim is Theorem 6.17: every symmetric fusion category over $\\mathbb{R}$ is equivalent to $\\mathrm{sRep}_{\\mathbb{C}/\\mathbb{R}}(G,z,s)$ for some finite super group $(G,z)$ with a group homomorphism $s:G\\to\\mathbb{Z}_2$ satisfying $s(z)=0$. A semi-linear super representation is a finite-dimensional complex super vector space $V$ with a $\\mathbb{C}$-linear or anti-linear action of $G$ according as $s(g)=0$ or $1$, such that $z$ acts by the parity operator, and the category is symmetric monoidal and real-linear rather than complex-linear in general. The reduction is by complexification: a real symmetric fusion category becomes a symmetric fusion category over $\\mathbb{C}$, which the complex classification (Theorem 3.11) identifies with $\\mathrm{sRep}_{\\mathbb{C}}(K,z)$; the remaining choice is a semi-linear $\\mathbb{Z}_2$-action, and Sections 6.1-6.3 show these actions correspond exactly to $\\mathbb{Z}_2$-graded extensions of $(K,z)$, with equivariantization producing $\\mathrm{sRep}_{\\mathbb{C}/\\mathbb{R}}(G,z,s)$. The paper further proves Theorem 6.22, an opposite equivalence between the 2-groupoid of real symmetric multi-fusion categories and finite groupoids equipped with a $\\mathbb{Z}_2\\times B\\mathbb{Z}_2$-action, realized by the groupoid of $\\mathbb{R}$-linear fiber functors into $\\mathrm{sVec}_{\\mathbb{C}}$.","pith_inferences":["An independent computation of the 2-group of semi-linear autoequivalences of $\\mathrm{sRep}_{\\mathbb{C}}(K,z)$ for small $(K,z)$, such as $\\mathrm{sVec}_{\\mathbb{C}}$ or $\\mathrm{Rep}_{\\mathbb{C}}(\\mathbb{Z}_2)$, would both test the omitted Proposition 3.15 and give explicit generators for the $\\mathbb{Z}_2\\times B\\mathbb{Z}_2$-action appearing in the reconstruction theorem.","The same descent framework, using the semi-linear $\\mathbb{Z}_2$-action machinery of Section 4, could compute real forms of non-symmetric fusion categories even though the complex classification of those categories is not known; the paper restricts attention to the symmetric case.","In physics terms, the grading $s$ records whether a symmetry is time-reversal-like and the central element $z$ records fermion parity, so Theorem 6.17 gives a normal form for any finite symmetry group of a fermionic system with anti-unitary symmetries; this application is not developed in the paper."],"forward_implications":["Every finite bosonic or fermionic symmetry of a quantum system over the real numbers appears as $\\mathrm{sRep}_{\\mathbb{C}/\\mathbb{R}}(G,z,s)$; there are no real symmetric fusion categories outside this family.","The explicit data are finite super groups with a $\\mathbb{Z}_2$-grading, so listing all real symmetric fusion categories is equivalent to listing $\\mathbb{Z}_2$-graded extensions of the finite super groups $(K,z)$ appearing over $\\mathbb{C}$.","Theorems 6.23 and 6.24 characterize Tannakian versus super-Tannakian real categories: they are $\\mathrm{Rep}_{\\mathbb{C}/\\mathbb{R}}(K\\rtimes\\mathbb{Z}_2^T)$ and $\\mathrm{sRep}_{\\mathbb{C}/\\mathbb{R}}((K,z)\\rtimes\\mathbb{Z}_2^T)$ respectively, with $K$ (resp. $(K,z)$) carrying a $\\mathbb{Z}_2$-action.","Theorem 6.22 says a real symmetric multi-fusion category and its concrete fiber-functor groupoid determine each other completely, giving a reconstruction result for real categories analogous to the complex Tannaka-Krein theorem.","Conjecture 6.26 proposes the same shape $\\mathrm{sRep}_{E/F}(G,z,s)$ for general characteristic-zero fields, with $E/F$ a finite Galois extension and $G$ graded by $\\mathrm{Gal}(E/F)$, so the method is expected to extend beyond $\\mathbb{C}/\\mathbb{R}$."],"supporting_citations":[{"why":"Supplies the classical classification over algebraically closed characteristic-zero fields, identifying the complexified category with $\\mathrm{sRep}_{\\mathbb{C}}(K,z)$.","marker":"[Del90, Del02]"},{"why":"Gives the fiber functor to $\\mathrm{sVec}$ and the finite group of monoidal automorphisms used to reconstruct the complex category from fiber functors.","marker":"[Del90, Section 8.19]"},{"why":"Establishes that every symmetric fusion category over an algebraically closed characteristic-zero field admits a fiber functor to $\\mathrm{sVec}$, the input to the reduction.","marker":"[Del02, Corollary 0.8]"},{"why":"Develops descent and forms of tensor categories; the $\\mathbb{C}/\\mathbb{R}$ special case becomes Theorem 4.14 turning real forms into semi-linear $\\mathbb{Z}_2$-actions.","marker":"[EG11]"},{"why":"Classifies Tannakian categories with fiber functors to $\\mathrm{Vec}$ and supplies the reconstruction principle underlying the Tannaka-Krein correspondence.","marker":"[DM82, Theorem 2.11]"},{"why":"Provides the connectivity of the fiber-functor groupoid used in Lemmas 6.1 and 6.3 and in Proposition 6.4 to control possible semi-linear actions.","marker":"[EGNO15, Theorem 9.9.26 (ii)]"},{"why":"Identifies the complexification of $\\mathrm{Rep}_{\\mathbb{C}/\\mathbb{R}}(G,s)$ with $\\mathrm{Rep}_{\\mathbb{C}}(\\ker s)$, anchoring the example class used throughout the classification.","marker":"[KZ24, Proposition* 6.11]"}],"fun_headline_variants":["Real symmetric fusion categories are super group semilinear reps","Classification over R: symmetric fusion categories = super group reps","Galois descent cracks real symmetric fusion categories","Tannaka-Krein for real: fusion categories as groupoid actions"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole classification rests on the assumption that every way of adding an anti-unitary symmetry to a complex super-representation category can be written, up to isomorphism, as complex conjugation followed by an ordinary unitary symmetry, a step the paper states as a proposition whose proof is omitted.","fun_headline_variants_meta":{"raw":{"variants":["Real symmetric fusion categories are super group semilinear reps","Classification over R: symmetric fusion categories = super group reps","Galois descent cracks real symmetric fusion categories","Tannaka-Krein for real: fusion categories as groupoid actions"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001259,"raw_usage":{"total_tokens":5169,"prompt_tokens":968,"completion_tokens":4201,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":584,"completion_tokens_details":{"reasoning_tokens":4135}},"tokens_in":584,"tokens_out":4201,"duration_ms":27441,"temperature":1.0,"reasoning_tokens":4135,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T14:38:04.681221+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute, for the super group $(K,z)=(\\mathbb{Z}_2,1)$, i.e. for $\\mathrm{sVec}_{\\mathbb{C}}$, all semi-linear $\\mathbb{Z}_2$-actions and their equivariantizations; the paper's classification predicts exactly two actions, corresponding to the extensions $\\mathbb{Z}_2\\times\\mathbb{Z}_2$ and $\\mathbb{Z}_4$, and a third action or a third equivariantization would contradict Theorem 6.17.","supporting_citations":[],"review_version":2}