{"id":"37c01b67-caa2-4444-b7da-a4dba0dbf169","arxiv_id":"2603.19006","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Analytic continuation of Nc in Yang-Mills theory produces non-Hermitian operator spectra with exceptional points, PT-phase transitions, and topological monodromy.","lead":"This paper shows that continuing the number of colors Nc to complex values makes the operator-mixing (dilatation) matrix of Yang-Mills theory non-Hermitian, with exceptional points where operator eigenvalues and eigenvectors coalesce. It argues these points produce PT-symmetry breaking, logarithmic scaling, and non-Abelian braiding, and may constrain the validity of the 1/Nc expansion.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The negative-norm states underpinning the EP are only shown to appear in a formally continued Gram matrix; no construction shows they are states of a non-integer-Nc Yang-Mills theory, so the EP/PT claims may be artifacts.","rationale":"The reader's weakest_assumption and my concern coincide: the analytic continuation of the operator inner product is the point where the argument is least secure. The one-loop EP is a concrete, reproducible matrix calculation and I do not dispute it; the issue is whether the objects being diagonalized belong to Yang-Mills at non-integer Nc. Deligne categories are a plausible framework, but the paper does not show how the interacting gauge theory—gauge fixing, Faddeev-Popov determinants, asymptotic states—is represented in them, nor that the cut-based Gram matrix equals the categorical inner product. The test I propose would provide the missing non-integer check. Because this is exactly the assumption the reader identified, and because the narrow spectral result is not affected by my critique, the conditional verdict stands: the mathematical EP claims can be accepted, but the physical interpretation requires further evidence. I therefore do not move the verdict.","tokens_in":28437,"tokens_out":17885,"duration_ms":207640,"concrete_test":"Independently compute the leading-order two-point functions of the eight dimension-8 length-4 operators at, say, Nc=2.9 and Nc=3.5 using the explicit Deligne-category trace rules from ref. [38] (treating the SU(Nc) generators as morphisms in Rep(GL_t) with t=Nc). Compare the resulting 8x8 Gram matrix with the rational expressions in Eq. (24) and the ancillary data. If they match at both non-integer points, the unitarity-cut derivation is validated as the categorical inner product. If they differ, Eq. (24) is not the inner product of the continued theory, and the negative-norm/EP interpretation is unsupported. A lighter complementary check: verify that a one-loop four-point function in this sector satisfies crossing symmetry as a rational function of Nc.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central one-loop EP at Nc=2.82466 is a well-defined property of the printed matrix D(1) (Eq. 29); my concern is the step before it. The paper identifies the leading-order two-point coefficient matrix G(0) (Eq. 24) with the inner product of the operator space, computing it through the cut formula (Eq. 12) with the SU(Nc) completeness relation (Eq. 14). For non-integer Nc this is a formal algebraic continuation: there is no Hilbert space over which the 'sum over physical states' in Eq. (12) runs, and Eq. (14) is an algebraic identity, not a resolution of the identity. Sec. VI appeals to Deligne categories (refs. [38,39,59]) but supplies no construction of the interacting gauge theory at complex Nc, no demonstration that the perturbative Gram matrix is the categorical inner product, and no check of consistency conditions (crossing, OPE associativity) at non-integer Nc. If the continuation is only a bookkeeping device, the negative-norm color-evanescent states are basis artifacts, and the non-Hermitian H, the EP, and the PT transition lose their stated physical meaning for Yang-Mills.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes that analytic continuation of the number of colors Nc in SU(Nc) Yang–Mills theory produces a non-Hermitian operator structure. Working with dimension-8, length-4 gauge-invariant operators, including color-evanescent operators, the authors compute leading-order Gram matrices and one- and two-loop dilatation matrices via on-shell unitarity. For the (−)^4 helicity sector, the Gram matrix (Eq. (24)) is indefinite for Nc < 3, and the one-loop dilatation matrix (Eq. (29)) has an exceptional point at Nc = 2.82466 where two eigenvalues and their eigenvectors coalesce (Eqs. (30), Figs. 5 and 6). The paper interprets this as a PT-symmetry phase transition, logarithmic scaling of correlators at the EP, and non-Abelian monodromy in the complex-Nc plane, and extends the analysis to dimension-evanescent sectors and higher-length operators.","tokens_in":28729,"tokens_out":7349,"duration_ms":77080,"significance":"If the complex-Nc continuation is legitimate, this is a concrete, parameter-free demonstration of exceptional points arising from within a unitary QFT rather than from an ad hoc non-Hermitian Lagrangian. The one-loop EP is obtained from explicit finite matrices, and the consistency condition D(1)·G(0) symmetric is a nontrivial internal check. The paper also makes falsifiable statements within its framework: complex anomalous dimensions, logarithmic oscillations, and monodromy of operator branches. Its main weakness is that the entire interpretation depends on the existence of a well-defined interacting gauge theory at non-integer Nc with the leading-order two-point matrix as its operator inner product; this foundation is asserted but not established.","major_comments":[{"comment":"The load-bearing assumption is that the leading-order two-point coefficient matrix G(0), computed from the cut sum in Eq. (12) with the completeness relation Eq. (14), is the inner product on the operator space for non-integer Nc. For non-integer Nc there is no Hilbert space of physical gluon states over which the sum in Eq. (12) runs; Eq. (14) is an algebraic identity, not a resolution of the identity. The appeal to Deligne categories in §VI does not supply a construction of the interacting gauge theory at complex Nc, nor a proof that G(0) equals the categorical inner product, nor a consistency check such as crossing or OPE associativity. Without this, the negative-norm states, and hence the EP, are properties of a formal algebraic extrapolation rather than of Yang–Mills theory. I would need at least one independent check at non-integer Nc, e.g., a three-point function computed by two d","section":"§III.A, Eqs. (12)–(14); §VI"},{"comment":"The 'emergent' PT symmetry is constructed by declaring P to be the signature matrix of G and T to be complex conjugation. For any real dilatation matrix D and any real symmetric G with signature P, the symmetrized matrix H = M D M^{-1} satisfies [PT,H]=0 whenever M^* = P M; this is a formal consequence of the indefinite metric, not a dynamical property inherited from spacetime PT. The identification with spacetime PT rests entirely on the i^L phase convention introduced in Appendix D, but the operator basis in Eq. (A1) is displayed without those phases, so Eqs. (41)–(43) do not establish that the spectral transition is 'spontaneous breaking of spacetime PT'. The authors should either prove that the phase convention is forced by the basis used in the Gram and dilatation matrix calculations, or present the PT correspondence as an interpretation rather than a derivation.","section":"§V.A, Eqs. (35)–(43) and Appendix D"},{"comment":"The persistence of EPs at two loops and the numerical shifts in Eq. (31) are central to the claim of NLO robustness, but they are given only as numbers. No two-loop 8×8 dilatation matrices, master-integral reductions, or consistency checks appear in the text or appendices; the reader is referred to ancillary files. Those files are not part of the refereed record. Since the paper advertises the first two-loop full-color calculation, omitting this material makes the NLO conclusion unverifiable. Please include the two-loop matrix data, or at least the discriminant/EP conditions, in the manuscript, or explicitly demote the two-loop statements to a preliminary remark.","section":"§IV.B, after Eq. (31)"}],"minor_comments":[{"comment":"The construction M = sqrt(P) R and the claimed property M^* = P M should be derived explicitly. For indefinite P the square root is non-unique, and the displayed H(1) contains many square-root factors that obscure the general argument.","section":"Appendix C, Eq. (C4)"},{"comment":"The column headings and the entries 'P T /x69+ − −' are garbled. Since the parity assignments of A_mu and ∂_mu are central to Eq. (43), they must be stated in unambiguous notation.","section":"Appendix D, Table II"},{"comment":"The entry '0 .231' appears to be a typo for 0.231; the matrix display also mixes decimal and symbolic entries, making the text hard to read.","section":"Appendix F, Eq. (F3)"},{"comment":"The notation 'α_s/4π' is used inconsistently with the surrounding text. The expansion parameter of the dilatation matrix should be defined once in §III B and used uniformly.","section":"§IV.B, Eq. (31)"},{"comment":"The labels λ1,...,λ4 are not explained in the caption, and the color convention for complex-conjugate pairs is only given in one panel. Please specify each curve and the phase-region boundary.","section":"Fig. 5"},{"comment":"There is an unresolved cross-reference 'Figure ??' in the discussion of the interval where the spectrum remains real.","section":"Appendix C"}],"recommendation":"major_revision","confidential_remarks":"The manuscript builds on the authors' earlier work on color-evanescent operators, and the genuinely new contribution is the EP/PT/monodromy interpretation. The explicit one-loop result is a solid finite-matrix computation, but the foundational status of non-integer Nc in an interacting gauge theory is the key risk. If the authors cannot provide independent consistency checks for the complex-Nc continuation, the advertised topological and PT claims should be substantially reframed as conjectures about a formal extrapolation."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"You should know this one before anything else: the one-loop exceptional point is real and checkable. The dilatation matrix is printed, the discriminant is given, and the eigenvector coalescence follows from the matrix. That part is solid. The broader physical interpretation — PT symmetry, monodromy, logarithmic CFT behavior — is more fragile, and the authors mostly know it. They flag the Deligne-category dependence in Sec. VI, but they don't construct the interacting theory at complex Nc.\n\nWhat is actually new: earlier papers by this group found negative-norm color-evanescent states and complex anomalous dimensions. Here the degeneracy is identified as an exceptional point, with eigenvector coalescence, a Jordan-block limit, and monodromy matrices. The two-loop full-color dimension-8 dilatation matrices are also new. The one-loop result is fully reproducible from the text: D^(1) is explicit, the discriminant is a polynomial in Nc^2, and the coalescence plot is a direct computation. That is real formal content and should be credited.\n\nSoft spots, in proportion: the load-bearing assumption is the analytic continuation itself. The paper identifies the leading-order Gram matrix with the operator inner product and treats the SU(Nc) completeness relation as a resolution of the identity. For non-integer Nc that is a formal algebraic continuation; no Hilbert space or categorical inner product for the interacting theory is actually constructed. So the negative-norm states could be basis artifacts, and if that is the case the EP loses its stated physical meaning. This is not a fatal flaw in the narrow claim — the EP is a genuine property of the printed matrix — but it does mean the ambitious claims about spacetime PT correspondence, topological constraints at Nc=3, and divergence of the planar expansion rest on something the paper has not shown. The effective PT symmetry is also partly built by definition: the parity operator is chosen to be the signature matrix of the Gram metric, so the correspondence to spacetime PT is more an observation about the construction than an independent derivation. Minor issues: the two-loop and length-5 data live only in ancillary files, and there is a broken figure reference near App. C.\n\nWho this is for: people working on non-Hermitian QFT, operator mixing at full color, or Deligne-category applications to gauge theory. It will not change how we think about real QCD at Nc=3, since the EP sits at Nc=2.825, but the explicit matrices are useful reference data and the questions about the planar expansion are worth taking seriously.\n\nRecommendation: send it to peer review. The core one-loop result deserves referee time. Ask the authors to include the ancillary data as supplementary material, fix the figure reference, and either justify the complex-Nc continuation more concretely or clearly separate the formal EP results from the interpretive PT/topology framework.","headline":"One-loop EP is explicit and reproducible from the printed matrices; the physical story at non-integer Nc is honest but underbuilt, so the paper deserves refereeing with a request to ship data and separate math from interpretation.","tokens_in":29224,"tokens_out":3080,"would_cite":true,"duration_ms":33123,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["81T13","81T15"],"pacs":["11.15.-q"],"model":"deepseek-v4-flash","headline":"The paper's central claim is that analytic continuation of Nc to complex values makes the Yang-Mills dilatation operator non-Hermitian, with an exceptional point at Nc=2.82466 where two operator eigenstates coalesce.","keywords":["Yang-Mills theory","analytic continuation of Nc","color-evanescent operators","exceptional points","non-Hermitian operator mixing","PT symmetry","anomalous dimensions","monodromy"],"falsifier":"Compute the one-loop Gram and dilatation matrices for the (−)4 dimension-8 sector by a method independent of on-shell unitarity (e.g., direct Feynman-diagram renormalization) and diagonalize D(1) at Nc=2.82466: if the two eigenvalues are not exactly degenerate and their eigenvectors do not become parallel, the EP is an artifact. A second check: continue the eigenvalues along a small closed contour around Nc=2.82466; if after one circuit the eigenvectors return to themselves rather than being permuted by the matrix in Eq. (55), the topological-monodromy claim fails.","tokens_in":28306,"feed_emoji":"🌀","tokens_out":8941,"duration_ms":84190,"temperature":0.7,"pith_summary":"This paper tries to establish that treating the number of colors, Nc, as a complex variable instead of an integer turns the operator spectrum of ordinary Yang-Mills theory into a non-Hermitian problem with exceptional points — degeneracies where both eigenvalues and eigenvectors collide. The driver is a class of color-evanescent operators that vanish at certain integer Nc but acquire negative norm in the continued theory. Because the two-point-function inner product becomes indefinite, the dilatation matrix (the effective Hamiltonian controlling operator mixing) becomes non-Hermitian, and pairs of anomalous dimensions can collide and go complex. The authors find a concrete collision at Nc=2.82466 in the dimension-8 length-4 sector, confirm it survives at two loops, and show that circling these points in the complex Nc plane swaps operator identities. A sympathetic reader would care because the result embeds unitary Yang-Mills in a larger complex parameter space with topological defects, without altering the Lagrangian.","feed_headline":"At Nc=2.82466, Yang-Mills operator spectrum hits an exceptional point","feed_subtitle":"At that point two operator eigenstates fuse; circling it in complex-Nc space swaps their identities.","key_machinery":"The load-bearing device is the color-evanescent operator: an operator built with a rank-n generalized Kronecker symbol that vanishes at integer Nc