{"id":"6f6d2bbc-20cc-4de3-8462-72a365d504ad","arxiv_id":"2608.00166","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":9,"one_line_summary":"A hand-imposed hierarchical texture for two flavour spurions fits the SM quark/lepton masses and CKM, and the same spurions make the electron EDM and K0-Kbar0 mixing the strongest probes of the SU(15) preon scale.","lead":"This paper shows that a proposed SU(15) composite model can be made to reproduce all Standard Model fermion masses and quark mixings, if its two flavour-breaking matrices are chosen by hand. It then uses the same matrices to predict which rare decays and electric-dipole searches would first reveal the compositeness scale, with the electron EDM and K0-Kbar0 mixing leading the way.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The hand-imposed κ-texture of Eq. (3.2) may be doing all the work: no scan over alternative integer textures is reported, so the 'reproduction' of masses and CKM could be generic rather than a prediction of the model.","rationale":"The reader's weakest_assumption identifies the hand-imposed texture as the load-bearing premise; my independent read agrees. The paper is an honest existence proof, but the existence claim is weakened by the lack of a texture scan and by the fact that the benchmark fit sits at the boundary of the allowed non-perturbative coefficient ranges. The proposed texture scan directly tests whether the model's successful mass/CKM reproduction is a non-trivial consequence of the SU(15)_p setup or an artifact of fitting freedom. Since the verdict is already CONDITIONAL and my concern reinforces the same conditionality, no change to the verdict is needed.","tokens_in":36216,"tokens_out":7753,"duration_ms":97559,"concrete_test":"Run the Appendix C minimisation on an ensemble of random integer-power textures: for each of ~10^3 choices of n_ij in Eq. (3.2) (respecting the antisymmetry/symmetry of λ/λ', with powers drawn uniformly from 0..6 and λ'44 fixed at 1), perform the same Adam/L-BFGS fit and record the best loss. If more than a few percent of random textures achieve sub-10% agreement on the nine fermion masses, nine CKM moduli, and J with |X|≤6 and |δ_X|≤1, then the 'judicious' texture is not load-bearing and the model's flavour predictions are generic; if none or almost none do, the texture is doing essential work and needs a UV mechanism.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is that the two-scalar SU(15)_p model reproduces the observed quark/lepton masses and CKM with O(1) non-perturbative coefficients. The load-bearing step is the 'judicious' Froggatt–Nielsen texture of Eq. (3.2): every entry of λ and λ' is assigned an integer power of κ≈0.17, and these powers are chosen to yield the observed hierarchies. No dynamics generates this texture, and the paper tests only this one choice. The fit itself is heavily underdetermined (~40 parameters vs ~14 observables), and the benchmark is a single point in a highly degenerate parameter space. Moreover, the fitted non-perturbative coefficients are not all O(1) in the usual sense: Appendix B.1 lists J_ℓ=+5.647, K_ℓ=−4.248, G_d≈4.9, with several up/down splittings δ_X at or near the imposed bound |δ_X|≤1 (δ_G=+1.000, δ_F=+0.962, δ_J=+0.913). Appendix B.2's claim that these are 'within a factor of ≲2 of the O(1) NDA expectation' is numerically inconsistent with values above 5. If a generic texture can also fit the data, then the specific texture is not a consequence of the model, and the FCNC reach numbers in Fig. 6 are not robust predictions of the SU(15)_p dynamics.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the flavour structure of the SU(15)_p confining chiral gauge theory proposed by Dobrescu, in which SM quarks, leptons and the Higgs are composite prebaryons. Two scalars in the conjugate antisymmetric (105) and conjugate symmetric (120) representations of SU(15)_p provide two SU(4)_F spurions, λ (antisymmetric) and λ′ (symmetric), which generate both the SM Yukawa couplings and flavour-changing processes. The up/down Yukawas are correlated by an approximate right-handed isospin symmetry, so a non-trivial CKM matrix requires spontaneous SU(2)_I breaking by composite-scalar vevs. Using a hand-imposed Froggatt–Nielsen-like κ-texture with κ≈0.17, the authors perform a numerical fit (order 40 free parameters) and present a benchmark that reproduces the six quark masses, three charged-lepton masses, and CKM matrix with coefficients claimed to be O(1). The same spurions are then used to compute dipole, semi-leptonic, four-quark and leptonic four-fermion operators, yielding constraints on the compositeness scale Λ_pre from e-EDM, K0–Kbar0 mixing, μ→eγ, μ→e conversion, D and B mixing, rare kaon decays, and related observables. The headline results are that e-EDM and ϵ_K probe Λ_pre near 10^4 TeV currently, with projected e-EDM sensitivity reaching ~10^6 TeV, beyond proton-decay projections.","tokens_in":36825,"tokens_out":5330,"duration_ms":65568,"significance":"If the benchmark construction is accepted, the paper is a useful first systematic study of flavour in this composite model. It gives an explicit, fully documented numerical benchmark (including an ancillary machine-readable file), a clear operator dictionary for low-energy flavour observables, and a broad comparison of current and projected experimental reaches. The logical connection between the mass fit and the FCNC/EDM predictions is non-circular: the fitted spurions are used to predict independent flavour observables. The paper is also commendably explicit about many of its own limitations. However, the central claim that the model 'reproduces' the fermion spectrum is weakened by three factors: the κ-texture is imposed rather than derived or statistically motivated; the fit is heavily underdetermined (~40 parameters vs 14 observables) and only one benchmark is shown; and the fitted non-perturbative coefficients are not all O(1) in the sense used in the abstract and conclusions, with values up to |X|≈5.6 and u/d splittings at the imposed bound. The FCNC reach numbers therefore inherit an unquantified theory uncertainty from the unknown non-perturbative coefficients, which the pa","major_comments":[{"comment":"The central 'reproduction' of quark and lepton masses and CKM rests on the hand-imposed κ-texture in Eq. (3.2). No dynamics is shown to produce these integer powers, and no scan over alternative textures is reported; the paper itself states in §4.6 and the Conclusions that other spurion choices 'would generically lead to new structures' and may be realized in the UV. With ~40 fitted parameters and 14 observables, the presented benchmark is a single point in a highly degenerate space. To make the claim 'the model reproduces the observed spectrum' load-bearing, the authors should either (i) perform a scan over integer κ-textures with the same κ and report the fraction that fit the data, or (ii) identify a symmetry or dynamical argument that uniquely selects Eq. (3.2). Without this, the mass/CKM 'prediction' is a fit to the observables it claims to reproduce.","section":"§3.1, Eq. (3.2); §C"},{"comment":"The claim that all fitted non-perturbative coefficients are O(1) is not supported by the benchmark values. Appendix B.1 gives J_l=+5.647, K_l=−4.248, and the u/d splittings include δ_G=+1.000, δ_F=+0.962, δ_J=+0.913, i.e. several at or near the imposed bound |δ_X|≤1. Appendix B.2 states these are 'within a factor of ≲2 of the O(1) NDA expectation', which is numerically incorrect for coefficients with |X|≈5.6. The abstract and conclusions repeat the O(1) claim. The authors should either redefine what 'O(1)' means in this paper (e.g., |X|<2π), or, if the intended NDA range is larger, state it explicitly and propagate it to the naturalness assessment of the benchmark.","section":"App. B.1, B.2; §5"},{"comment":"First-generation observables are reproduced only through cancellations. The text after Eq. (3.17) states that m_u requires cancellations 'on the order of a few×10^−2', and Appendix B.2 reports that m_u, m_e, |V_ub|, |V_td| and the Jarlskog invariant J have large multiplicative responses to a 3% jitter of the fitted parameters. These are precisely the observables that differentiate the texture from a generic hierarchical ansatz. The claim 'all of the fitted non-perturbative coefficients are O(1)' therefore hides a significant fine-tuning hotspot. A quantitative fine-tuning measure (e.g., Barbieri–Giudice-type sensitivity) should be reported for the benchmark, and the abstract/conclusions should be adjusted so that 'O(1) coefficients' is not read as 'natural reproduction of all masses and CKM elements.'","section":"§3.1 (after Eq. 3.17); App. B.2"},{"comment":"The flavour-physics reach estimates in Fig. 6 are obtained by setting all non-perturbative coefficients in the dipole and four-fermion operators to 1, while the mass fit in App. B.1 finds coefficients spanning roughly 0.4 to 5.6. Since the bounds scale as a square root (or fourth root) of the Wilson coefficients, the quoted Λ_pre values can shift by factors of order 2–3, and the ordering of the most sensitive observables could change. The paper acknowledges this caveat in §4.6, but the abstract and Fig. 6 present the reach numbers as definite. The authors should propagate the fitted coefficient range into the benchmark bounds, or at least show a band of Λ_pre for each observable corresponding to the spread found in App. B.","section":"§4.2, Eq. (4.3); §4.3–4.6"}],"minor_comments":[{"comment":"In the ∆F=2 Lagrangian, the term written as [C^{ℓq}_{RL}]_{ji;ij} should presumably be [C^{qq}_{RL}]_{ji;ij} (or an analogous quark–quark coefficient); as written it mixes lepton and quark labels in a four-quark operator.","section":"Eq. (4.60)"},{"comment":"The sentence about pNGBs lying 'two orders of magnitude [taken from sqrt(...)]' appears corrupted or incomplete; the parenthetical material should be integrated into the text.","section":"§3.3"},{"comment":"The statement 'within a factor of ≲2 of the O(1) NDA expectation' should be reworded once the criterion for O(1) is defined; as written it is inconsistent with the quoted values.","section":"App. B.2"},{"comment":"The abstract's 'O(1) non-perturbative coefficients' is too strong given the fitted values in App. B; consider saying 'within an order of magnitude of unity' and referencing the residual fine-tuning in the first generation.","section":"Abstract / §5"},{"comment":"The caption is dense; it would help to state explicitly that the solid/dashed bars correspond to current/projected experimental sensitivity and that the benchmark panel uses cosβ=0.0437 while the anarchic panel uses cosβ=1.","section":"Fig. 6 caption"}],"recommendation":"major_revision","confidential_remarks":"This is a technically serious exploratory paper with a useful and clearly documented benchmark. My main reservation is that the head-line claim of 'reproducing' the fermion spectrum is a fit over a hand-imposed texture with ~40 parameters, and the paper does not quantify how generic this success is. The O(1) claim is also internally inconsistent with App. B. These are fixable in revision by adding a texture scan or symmetry justification, propagating the fitted coefficient spread into the FCNC bounds, and tempering the abstract. I do not think rejection is warranted: the phenomenological framework and the FCNC–mass connection are valuable, and the paper is unusually transparent about its assumptions."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear —,\n\nQuick take: this is the first real flavour study of the two-scalar SU(15) preon model. It builds the Yukawa operator basis, corrects an earlier 105 vs 120 representation mistake, and works out which low-energy observables probe the compositeness scale. The authors are unusually honest about what is assumption and what is computation.\n\nWhat is genuinely useful: the spurion construction and the phenomenological survey. The e-EDM and ε_K reaching ~10^4 TeV, with projected e-EDM out to 10^6 TeV, is a sharp, concrete statement that flavour can compete with proton decay for this model. The benchmark is explicit, with machine-readable file, and the robustness jitter in App B.2 is good practice. They also flag that the UV texture is not derived and that the nonperturbative coefficients are placeholders.\n\nSoft spots, in proportion. The central 'reproduces masses and CKM' is a ~40-parameter fit into 14 observables. The κ-texture is imposed by hand — every power is chosen to generate the observed hierarchy. No dynamics produces it, and no scan over alternative integer textures is shown. If another texture fits equally well, the FCNC reach numbers in Fig. 6 are not robust predictions of the SU(15)_p dynamics. The stress-test note is right about this; it is not fatal for an existence proof, but it should be said more prominently than the abstract does.\n\nSecond, the 'O(1) coefficients' claim is stretched. App B.1 has J_ℓ = 5.647, K_ℓ = −4.248, G_d ≈ 4.9, and several δ_X pinned at the boundary ±1. The text says these are within a factor of ~2 of the NDA O(1) expectation. Factor ~6 is not factor ~2. That overstatement should be fixed.\n\nThird, the matching assumes the strongly-coupled chiral gauge theory is captured by a few tree/one-loop topologies with O(1) coefficients. There is no lattice control. The authors acknowledge this in Sec 2.4, so I don't count it as hidden; it just caps how much weight the reach numbers can carry.\n\nBottom line: this deserves a serious referee. With the texture caveat and the coefficient-size overstatement addressed, it is a useful existence proof and a road map for flavour probes of composite models. I would send it to review, but ask for a texture genericity scan, error propagation of the O(1) coefficients, and a corrected O(1) claim.\n\nBest","headline":"Honest and careful first flavour analysis of the two-scalar SU(15) preon model, but the headline 'reproduction' is a texture-fit existence proof rather than a derivation, and the 'O(1)' claim is stretched by fitted coefficients around 5–6.","tokens_in":37268,"tokens_out":2656,"would_cite":true,"duration_ms":31271,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper argues that an SU(15)_p confining preon model, with composite quarks, leptons, and Higgs, can account for the complete observed pattern of charged-fermion masses and CKM mixing using two flavour spurions with a hierarchical κ≈0.17","keywords":["preon compositeness","SU(15) chiral gauge theory","flavour spurions","Froggatt-Nielsen texture","fermion mass hierarchy","CKM matrix","flavour-changing neutral currents","electric dipole moments"],"falsifier":"Measure the electron EDM below about 10^-31 e cm: the paper's benchmark gives Λ_pre > 6×10^4 TeV when d_e < 10^-31 e cm, so a null result combined with independent evidence that Λ_pre ≈ 10^4 TeV (e.g., from proton-decay limits) would force the diagonal dipole coefficient below its O(κ³) texture value and falsify the benchmark.","tokens_in":36071,"feed_emoji":"⚛️","tokens_out":8133,"duration_ms":86128,"temperature":0.7,"pith_summary":"The paper's central thesis: an SU(15)_p confining gauge theory, in which quarks, leptons and the Higgs are composite prebaryon bound states, can generate the entire observed charged-fermion mass spectrum and the CKM matrix from just two SU(4)_F flavour-breaking spurions λ and λ′, provided these have a hierarchical Froggatt-Nielsen-like texture with small parameter κ≈0.17. The authors exhibit a benchmark fit in which all non-perturbative coefficients are order one and all six quark masses, three charged lepton masses, and the full CKM matrix are reproduced. Because the very same spurions generate flavour-changing neutral currents and electric dipole moments, the compositeness scale Λ_pre becomes testable in precision experiments: electron EDM and K0–K̄0 mixing currently probe Λ_pre near 10^4 TeV, and projected electron-EDM sensitivity raises the reach to about 10^6 TeV, beyond planned proton-decay searches. If the construction is right, flavour physics—not just proton decay—becomes a discovery channel for quark and lepton compositeness.","feed_headline":"Two flavour spurions fit all quark and lepton masses","feed_subtitle":"The same couplings set the electron EDM and kaon mixing, probing the preon scale up to 10^6 TeV.","key_machinery":"The key objects are the two flavour spurions λ (antisymmetric, transforming as a 6 of SU(4)_F) and λ′ (symmetric, a 10), generated by Yukawa couplings of the SU(15)_p scalars in the conjugate antisymmetric (105) and conjugate symmetric (120) representations respectively. The paper imposes a hierarchical texture on them, Eq. (3.2), with integer powers of κ≈0.17; this texture seeds the five-order-of-magnitude hierarchy of fermion masses. The machinery then consists of the 1/N and N/16π² power-counting rules: tree-level s-channel exchanges plus one-loop crossed-box dressings build the Yukawa matrices, while the misalignment between λ and λ′ (relative to the right-handed-isospin symmetric limit)","core_discovery":"On its own terms, the model's central claim is that a two-spurion flavour sector suffices. The antisymmetric λ and symmetric λ′ Yukawa couplings of the SU(15)_p scalars A and A′ break SU(4)_F explicitly, and with the κ-texture of Eq. (3.2) they produce up- and down-Yukawa matrices of the form F-term plus one-loop box structures weighted by N/16π². A numerical fit with coefficients bounded by |X|≲6, splittings |δ_X|≤1, and tanβ≈22.9 returns the observed m_u, m_d, m_s, m_c, m_b, m_t, m_e, m_μ, m_τ, the modulus of every V_CKM entry, and the Jarlskog invariant to sub-percent accuracy. The paper then derives the phenomenological consequence: rotating the same spurions to the mass basis induces di","pith_inferences":["The κ-texture is an input, not a prediction: the paper shows that if the UV completion delivers these powers, the fit works; finding a dynamical origin for κ and the exponents would convert a benchmark into a genuine prediction.","A null electron EDM at the projected 10^-34 e cm level would do more than push Λ_pre upward: it would force the diagonal dipole coefficient Im[Cγ_ℓ]11 below its O(κ³) benchmark value, ruling out the texture unless non-perturbative coefficients are tuned—providing a sharp discriminator between the hierarchical benchmark and flavour anarchy.","The same spurion logic could be exported to the neutrino sector, which the paper leaves to type-I seesaw: if the PMNS matrix were correlated with the charged-lepton rotations from λ, λ′ rather than anarchic, neutrino-oscillation data would indirectly test the preon scale.","The light composite states that mediate the lepton Yukawas (H_ℓ) could appear at colliders if Λ_pre is at the low end, so the flavour bounds and direct searches for TeV-scale scalars jointly constrain the model."],"forward_implications":["The compositeness scale Λ_pre must be at least about 10^4 TeV in the benchmark, since both the electron EDM and CP-violating K–K̄ mixing (ε_K) exclude lower scales; this matches the lower end of the proton-decay bound.","A next-generation electron EDM measurement at d_e ≲ 10^-34 e cm would probe Λ_pre ≈ 2×10^6 TeV, exceeding the planned Hyper-Kamiokande proton-decay reach.","µ→eγ, D–D̄ mixing, neutron EDM, B_d–B̄_d mixing and dipole-dominated µ→e conversion give complementary reach in the 10^2–10^3 TeV range, so a broad flavour program can map out the spurion texture.","The texture predicts specific parametric ratios—e.g., up/down Yukawa matrices nearly aligned, m_t/m_b set mostly by tanβ, and b→sγ, b→dγ both O(κ²)—that are testable in B-physics.","First-generation masses (m_u, m_e), small CKM elements |V_ub|, |V_td|, and the Jarlskog invariant require cancellations at the few-percent level, creating a localized fine-tuning hotspot that sharper measurements would stress."],"fun_headline_variants":["Two spurions fit all quark and lepton masses","Electron EDM probes preon scale to 10^6 TeV","SU(15) preon model: two spurions, one fit","Kaon mixing and EDM set preon scale beyond proton decay","Composite model matches masses and CKM with two flavour spurions"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"Everything rests on the hand-imposed κ≈0.17 texture of the two spurion matrices, together with the assumption that the strongly-coupled SU(15)_p dynamics is well described by a small set of O(1) matching coefficients; nothing in the model generates that texture, so a different UV pattern would shift every mass/CKM and FCNC prediction.","fun_headline_variants_meta":{"raw":{"variants":["Two spurions fit all quark and lepton masses","Electron EDM probes preon scale to 10^6 TeV","SU(15) preon model: two spurions, one fit","Kaon mixing and EDM set preon scale beyond proton decay","Composite model matches masses and CKM with two flavour spurions"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000397,"raw_usage":{"total_tokens":2021,"prompt_tokens":958,"completion_tokens":1063,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":702,"completion_tokens_details":{"reasoning_tokens":973}},"tokens_in":702,"tokens_out":1063,"duration_ms":10611,"temperature":1.0,"reasoning_tokens":973,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-04T01:07:50.262429+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the electron EDM below about 10^-31 e cm: the paper's benchmark gives Λ_pre > 6×10^4 TeV when d_e < 10^-31 e cm, so a null result combined with independent evidence that Λ_pre ≈ 10^4 TeV (e.g., from proton-decay limits) would force the diagonal dipole coefficient below its O(κ³) texture value and falsify the benchmark.","supporting_citations":[],"review_version":1}