{"id":"b133f706-a83e-4404-bb41-fc08317a51b4","arxiv_id":"2607.24901","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The natural functor from geometry-first generating objects to principal bundles with matter is faithful but neither essentially surjective nor full, so the two formulations of gauge theory are not categorically equivalent.","lead":"Geometry-first and symmetry-first gauge theory are not equivalent: a natural functor between them is faithful but neither essentially surjective nor full. The result matters for anyone who treats principal-bundle and vector-bundle formulations as notational variants.","discovery_kind":"extension","skeptic_critique":{"model":"moonshotai/kimi-k3","headline":"The proofs of Propositions 1–3 check out internally; the residual load-bearing question is whether non-equivalence survives arbitrary menu enlargements for compact groups, and that now rests almost entirely on the Appendix B \"order-two bound\" against triality-type outer automorphisms, which is asser","rationale":"The reader's weakest_assumption — menu-relativity of the classical-menu witnesses and the restriction of VBgen morphisms to strictly structure-preserving maps on matching fundamental fibres — is the correct place to press, and I found nothing more load-bearing. My independent pass over the mathematics (character lattice/algebraicity argument in Prop 2 and Appendix A; centre and π₁ separations for G' in Prop 1; the outer-automorphism, inner-shift, and normaliser computations NGL(SO(2m)) = R>0·O(2m) in Prop 3) found each step standard and correctly applied, consistent with the reader's HIGH confidence and medium correctness risk. I therefore agree with CONDITIONAL rather than ACCEPT: the two qualifications the reader cites (dependence on unpublished companions, especially Gomes & Weatherall 2026 for the splitting presentation of G', and menu-relativity) are real. My contribution is to localise the residual risk more precisely: because the compact witness is conceded to be menu-relative and the menu-independent witness is non-compact, the physically significant content of the non-equivalence claim is carried by the fullness failure, and its robustness under all source enlargements hinges on the Appendix B order-two bound against outer classes of order ≥ 3. That bound is proved in the text only for the specific real/complex scalar presentations at hand; its generalisation to arbitrary Chevalley-type tensorial presentations is plausible (the same tensors that cut the stabiliser down also constrain morphisms) but not demonstrated. The proposed Spin(8)/triality test is the cheapest decisive check: triality is the canonical order-3 outer automorphism, the paper itself invokes it as the limiting case, and a single explicit computation either confirms the bound in the hardest instance or exhibits a counterexample that would demote the central claim to a statement about one particular choice of category. Pending that check, CONDITIONAL with HIGH confidence is the right verdict, unchanged.","tokens_in":21551,"tokens_out":5601,"duration_ms":284429,"concrete_test":"Settle the menu-independent compact fullness obstruction directly. (a) Exhibit an explicit finite tensorial presentation of Spin(8): fibre so(8) with bracket plus the extra tensors (guaranteed by Chevalley/Deligne–Milne) whose full stabiliser is exactly Spin(8); take the trivial flat vacuum object over S¹. (b) Confirm the PFBmat automorphism (Φ,η,u) with η = triality exists (it does: Φ(x,g)=(x,η(g)) on the trivial bundle). (c) Enumerate all invertible structure-compatible semilinear self-maps of that presentation and compute their induced classes in Out(Spin(8)) ≅ S₃. If every such map induces an inner or order-2 class, the order-two bound holds in the critical case and the reader's menu-relativity concern is contained; if any presentation/morphism pair induces a triality class, the last menu-independent compact obstruction collapses and the verdict should move toward REJECT of the robus","verdict_should_be":"UNCHANGED","load_bearing_attack":"I re-checked the three propositions from the text and found no internal error. Prop 2 (additive R): the classification of connected 1-dimensional real linear algebraic groups (Ga, Gm, SO(2)), the full-stabiliser clause excluding R>0 ⊂ R×, and the degree argument χ(2t)=χ(t)² forcing χ≡1 are all correct, and the intertwining by a Lie group isomorphism η (necessarily a rescaling of R) does not rescue a nontrivial algebraic character. Prop 1's compact witness: the centre computation Z(G') = {(α1₃,β1₂): α³β²=1} ≅ U(1) (connected since gcd(3,2)=1) versus the disconnected centres U(1)×Z₆, U(1)×Z₃, U(1)×Z₂ of the candidate products, plus the π₁ exclusion of SO(3) variants, is sound. Prop 3: Ad_R is outer by the commutant-plus-determinant argument (λ^{2m} = −1 has no real solution for even rank), the trivial-flat (Φ,η,u) is a legitimate PFBmat automorphism, and the inner-shift stays within the outer class. So the stated claims are supported as stated.\n\nThe load-bearing exposure is scope, not correctness — and it is the same one the reader flagged, sharpened. The compact witness (G/Z₆) is admittedly menu-relative: a fixed 3⊕2 splitting gives G' a preimage. The menu-independent witness (charged additive R) is non-compact and physically exotic. Therefore everything physically consequential in the non-equivalence claim reduces to the fullness failure, and the only argument that fullness failure survives arbitrary enlargement of the source (menus and morphisms) is the Appendix B order-two bound: at any fixed presentation, structure-compatible semilinear maps realise at most one outer class, of order ≤ 2, so Spin(8) triality (order 3) can never be implemented source-side. That bound is proved only for the real/complex scalar-field setups \"used here\"; its extension to arbitrary finite tensorial presentations (e.g., Chevalley-type presentations of Spin(8) with extra tensors cutting Aut(so(8)) down to Spin(8)) is asserted. There is a self-correcting feature — any tensor added to shri","agreement_with_reader":"agree"},"referee_report":{"model":"moonshotai/kimi-k3","summary":"The paper compares two formulations of gauge theory: the standard symmetry-first (principal fibre bundle) picture and a geometry-first picture in which gauge groups arise as full stabilisers of structured fundamental vector bundles. It argues the two are not equivalent, in three senses: (i) the geometry-first theory-space is strictly smaller (charged additive-R theories have no finite tensorial presentation, Proposition 2); (ii) matter bundles do not determine the geometry-first provenance of the gauge group (the Standard Model's diagonal Z6 kernel: the faithful quotient G/Z6 is not a product of classical-menu stabilisers, §4.3); and (iii) the natural functor F : VBgen → PFBmat is faithful but neither essentially surjective nor full (Propositions 1 and 3), with fullness failing at the real oriented fibre R^{2m} because the outer automorphism Ad_R induced by an improper orthogonal map has no structure-preserving preimage. Appendix A supplies the algebraic-group facts behind Proposition 2; Appendix B analyses a complexified off-menu presentation and proves an order-two bound on outer classes realisable by semilinear maps, implying the triality automorphism of Spin(8) remains a fullness obstruction under menu enlargement.","tokens_in":21890,"tokens_out":2993,"duration_ms":142995,"significance":"If correct, this settles a question left open by the geometry-first programme (Gomes 2026a): whether that formulation is a notational variant of the principal-bundle formalism or a genuinely different structure. The answer given here is precise and falsifiable in form: the non-equivalence is witnessed by explicit constructions (the G/Z6 centre and π1 computations, the additive-R character obstruction, the concrete Ad_R automorphism on the trivial flat SO(2m) bundle), uses no fitted parameters, and the author is unusually candid about which witnesses are menu-relative (G/Z6) and which are not (additive R). The categorical framing (faithful, not full, not essentially surjective) engages the theoretical-equivalence literature (Weatherall, Barrett–Halvorson) on its own terms, and the observation that F's non-fullness blocks the standard \"restrict to the essential image\" response is a genuinely useful point. The derivation that incommensurate charge spectra distinguish the formulations modally (which future spectra each can represent) gives the result physical rather than merely classificatory content.","major_comments":[{"comment":"A morphism f : X → X' is defined only when (Va, structa) = (V'a, struct'a) for each a — i.e. literal equality of structured fibres, not isomorphism. This makes isomorphism classes in VBgen extremely fine-grained and directly controls the Hom-sets on which the fullness verdict depends (e.g. the empty Hom-set between XR and XC in Appendix B relies on a real-dimension mismatch, which is legitimate, but the general clause is stronger than needed). Please clarify whether equality or existence of a structure-preserving linear identification of fibres is intended, and confirm that the faithfulness and non-fullness arguments are insensitive to the choice. As written, a reader cannot tell whether VBgen is a category of presentations (fine) or of geometric objects (in which case the clause looks like a definitional artifact that stacks the deck against fullness).","section":"§5, definition of VBgen morphisms"},{"comment":"The claim that the fullness failure survives menu enlargement rests on the order-two bound: at any fixed presentation, structure-compatible semilinear maps realise at most one nontrivial outer class, so triality (order three) can never be implemented once Spin(8) is presentable. The argument given is correct for semilinear enlargements over R and C, but it bounds only that enlargement class. The conclusion 'no semilinear enlargement can realise an outer automorphism of order three' should be matched by an explicit scope statement in §5/§6: the non-fullness result is robust against enlarging the tensorial menu and against semilinear morphisms, but not against enlarging the source category to admit non-structure-preserving re-identifications (which the paper rightly notes would change the source-side notion of sameness rather than make F full). Since the compact essential-surjectivity witn","section":"Appendix B, order-two bound; end of §5"}],"minor_comments":[{"comment":"The sentence 'the VB-POV applies only to theories whose matter sector is generated tensorially from appropriate fundamental bundles, which excludes PFB-POV theories with non-compact gauge groups' overstates the result: §3 explicitly notes that non-compact groups (C×, GL(n,R)) are VB-presentable, and Proposition 2 excludes only additive R with charged matter. Suggest 'which excludes certain PFB-POV theories with non-compact structure group, such as charged additive-R theories'.","section":"§6, Conclusion"},{"comment":"Aharony, Seiberg & Tachikawa (2013) and Weatherall (2016c) appear in the reference list but I could not locate citations of them in the text; either cite (the former is natural near footnote 4 or the H^2(M,Z6) remark in footnote 5) or remove.","section":"References"},{"comment":"Typo: 'empirically commited' should be 'empirically committed'. In §6: 'Second ,they' has a misplaced space.","section":"§3"},{"comment":"The PSU(3) witness is a nice independent blocking mechanism, but the off-menu presentation via (bracket, cubic d-form) is compressed into one sentence; a line explaining why fixing d cuts Aut(su(3)) ≅ PSU(3) ⋊ Z2 to PSU(3) (i.e. that the outer automorphism flips the sign of d) would help readers not fluent in su(3) invariant theory. The parenthetical does say this; consider promoting it to the main text.","section":"§5, footnote 9"},{"comment":"The scalar-map shortcut (id, id, λ1) is mentioned after the main proof; it would help to flag before the proof that the reflection witness is chosen deliberately because it survives the strengthened (inner-product-preserving) target morphisms, so the reader does not wonder why the simpler witness was not used.","section":"§5, Proposition 3 and following"},{"comment":"The notation G ≃ ρ(G) ≃ Aut(V) is slightly abused later: in §4.3 the kernel discussion requires distinguishing G ≅ ρ(G) (faithfulness on the fibre) from faithfulness of the total matter representation ρ_tot. The text is aware of this, but a sentence at (2.5) noting that 'faithful' there means faithful on the fundamental fibre, not on matter, would prevent confusion.","section":"§2, Eq. (2.5)"}],"recommendation":"minor_revision","confidential_remarks":"The manuscript leans heavily on companion manuscripts (Gomes 2026a,b; Gomes & Weatherall 2026), several listed as 'Manuscript' or 'in preparation'; the load-bearing claims here are self-contained, but the editor may want to confirm the cited companion results (e.g. the S(U(3)×U(2)) presentation with fixed splitting) will be publicly available. The stress-test concern about the Appendix B order-two bound does land as a scope issue but not as a correctness issue — the bound as stated is proved adequately for semilinear enlargements; what is missing is an explicit robustness-class statement, which is a revision-level fix."},"author_rebuttal":null,"desk_editor":{"model":"grok-4.5","letter":"The new package is the functor F from geometry-first generating objects to principal bundles with matter, shown faithful but neither essentially surjective nor full, with three explicit witnesses: additive-R charged theories (every finite tensorial menu), the SM diagonal quotient G/Z6 on the classical menu, and the outer Ad_R automorphism for oriented R^{2m} recovering SO(2m).\n\nWhat works: Propositions 1–3 are clean and checkable from the text. Appendix A’s algebraic-character argument for additive R is standard and correctly applied (full stabiliser, no nontrivial algebraic characters). The G' centre and π1 separation from classical-menu products is right. The SO(2m) fullness failure is a straightforward outer-automorphism construction, and Appendix B is honest about semilinear enlargements, complexification, and the order-two bound. The paper does not hide that G' becomes presentable once a fixed 3⊕2 splitting is allowed, or that exceptional groups cost explanatory economy rather than presentability. Theory-space inclusion, recovery failure, and categorical non-equivalence are kept distinct. Citations to Weatherall/Barrett–Halvorson and the geometry are appropriate; companion self-cites are scaffolding, not load-bearing for the three propositions.\n\nSoft spots in proportion: the physically salient compact witness is menu-relative, and the menu-independent witness is non-compact and exotic. Fullness is what carries weight for ordinary compact groups; the order-two bound against triality-type outers is proved only for the real/complex scalar setups used here, and its extension to arbitrary Chevalley-type presentations is asserted rather than fully general. That is a scope limit, not an internal error. Dependence on unpublished companions for the broader geometry-first program is real but does not undercut the math in this paper.\n\nThis is for people who care about formulation equivalence and global form of gauge groups in foundations of gauge theory. The math is solid enough for a serious referee. I would engage and would send it out.","headline":"Solid categorical non-equivalence result with checkable witnesses; menu-relativity is flagged honestly and does not sink the stated claims.","tokens_in":23099,"tokens_out":506,"would_cite":true,"duration_ms":10048,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"grok-4.5","headline":"Geometry-first and symmetry-first gauge theory are not equivalent: fewer theories, unrecovered generators, and a non-equivalence of categories.","keywords":["gauge theory","geometry-first","symmetry-first","principal bundles","vector bundles","theoretical equivalence","Standard Model gauge group","categorical equivalence"],"falsifier":"Exhibit a finite list of tensors on a vector space whose full automorphism group is Lie-isomorphic to additive R and that carries a nontrivial algebraic character matching a charged representation, or exhibit a structure-preserving fibre map of the oriented Euclidean R^{2m} object that induces the outer automorphism Ad_R of SO(2m).","tokens_in":22684,"feed_emoji":"⚖️","tokens_out":928,"duration_ms":16041,"temperature":0.7,"pith_summary":"This paper argues that starting gauge theory from structured fundamental vector bundles (geometry-first) is not the same as starting from an abstract structure group on a principal bundle (symmetry-first). When the geometry-first side is limited to finite classical tensor data—inner products, volume forms, and the like—its automorphism groups cannot produce every symmetry-first theory; charged theories whose structure group is additive real numbers are permanently missing. Even when a theory does have a geometry-first presentation, the principal bundle and its matter fields do not tell you which fundamental bundles generated the gauge group; the Standard Model’s diagonal quotient is a concrete witness. Categorically, the natural map from geometry-first generating objects to principal bundles with matter is faithful but neither essentially surjective nor full, so it is not an equivalence. A sympathetic reader cares because the choice of primitive objects changes which theories are allowed, what can be recovered from matter, and the explanatory order of the gauge group itself.","feed_headline":"Geometry-first gauge theory is not the same as symmetry-first","feed_subtitle":"Fewer theories, unrecovered generators, and a functor that is faithful but neither full nor essentially surjective","key_machinery":"The functor F from VBgen (tuples of structured fundamental vector bundles with tensorial matter constructions) to PFBmat (principal bundles with connection and matter representations). It is shown faithful by frame-bundle projection, not essentially surjective by the Standard Model quotient G/Z6 on the classical menu and by additive-R theories on every finite tensorial menu, and not full because no structure-preserving fibre map induces the outer automorphism of SO(2m) coming from an improper orthogonal map.","core_discovery":"The geometry-first and symmetry-first formulations of gauge theory are not equivalent. They differ in the theories they admit (additive-R charged theories have no finite tensorial geometry-first presentation), in what matter bundles recover (they do not determine the generating structured bundles or the product-versus-quotient provenance of the group), and in categorical structure: the natural functor from geometry-first generating objects to principal bundles with matter is faithful but neither essentially surjective nor full.","pith_inferences":["Textbook treatments that treat principal-bundle and frame-bundle presentations as interchangeable for classical groups are silently assuming the geometry-first restrictions the paper makes explicit.","If future charge measurements ever produced ratios incommensurate with the known hypercharge lattice, every finite tensorial geometry-first presentation would be falsified at once while a symmetry-first R theory could absorb the new character.","The same normaliser/centraliser gap that blocks fullness for SO(2m) reappears whenever one tries to reconstruct a metric or volume form from transition data alone."],"forward_implications":["Geometry-first theory-space is a strict subset of symmetry-first theory-space once generators are restricted to finite tensorial data.","Matter bundles plus connection underdetermine whether the Standard Model group is the product SU(3)×SU(2)×U(1) or its faithful quotient by Z6.","The interaction-sector decomposition (which fundamental bundle each particle species shares) is extra structure not recoverable from the principal-bundle side alone.","No finite tensorial enlargement of the source can restore essential surjectivity while additive-R charged theories remain in the target.","Restricting the target to the essential image would change the modal empirical content of the framework, not merely its bookkeeping."],"fun_headline_variants":["Geometry-first gauge theory is not equivalent to symmetry-first","Geometry-first admits fewer gauge theories than symmetry-first","Additive-R charged theories lack finite tensorial geometry-first form","Functor from geometry-first objects is faithful but neither full nor ess. surjective","Principal bundles fail to recover geometry-first generating structures"],"cache_read_input_tokens":16512,"weakest_assumption_plain":"The comparison fixes a finite classical menu of fibre tensors whose full stabilisers define the geometry-first objects, and allows only structure-preserving maps of those fibres as morphisms; enlarging the menu or the allowed maps removes several of the witnesses.","fun_headline_variants_meta":{"raw":{"variants":["Geometry-first gauge theory is not equivalent to symmetry-first","Geometry-first admits fewer gauge theories than symmetry-first","Additive-R charged theories lack finite tensorial geometry-first form","Functor from geometry-first objects is faithful but neither full nor ess. surjective","Principal bundles fail to recover geometry-first generating structures"]},"model":"grok-4.5","effort":"low","cost_usd":0.004123,"raw_usage":{"total_tokens":1279,"prompt_tokens":779,"num_sources_used":0,"completion_tokens":71,"cost_in_usd_ticks":41228000,"prompt_tokens_details":{"text_tokens":779,"audio_tokens":0,"image_tokens":0,"cached_tokens":256},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":429,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":779,"tokens_out":71,"duration_ms":8348,"temperature":1.0,"reasoning_tokens":429,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-31T08:44:27.992576+00:00","model_set":{"reader":"grok-4.5"},"falsifier":"Exhibit a finite list of tensors on a vector space whose full automorphism group is Lie-isomorphic to additive R and that carries a nontrivial algebraic character matching a charged representation, or exhibit a structure-preserving fibre map of the oriented Euclidean R^{2m} object that induces the outer automorphism Ad_R of SO(2m).","supporting_citations":[],"review_version":1}