{"id":"7cfeda06-f4ce-4d26-b799-cffbadcb0941","arxiv_id":"2607.23497","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Vacuum-polarization flavor mixing in the U(3) NJL model produces small, mass-difference-driven shifts in charged pion and kaon valence PDFs that mildly improve agreement with data at 4 and 27 GeV².","lead":"A quark-model calculation finds that vacuum-polarization flavor mixing slightly reshapes pion and kaon valence PDFs, with the biggest shifts near x~0.3 and x~0.8. The shifts are small but may matter for precision meson-structure measurements planned at EIC and related facilities.","discovery_kind":"extension","skeptic_critique":{"model":"moonshotai/kimi-k3","headline":"The advertised residual magnitudes and peak locations depend on an arbitrary coupling-normalization convention (Eq. 16, fixing G11 = G0), which the paper itself flags as one choice among several; the quantitative headline may not be prescription-invariant.","rationale":"The reader identified the no-refit problem (mπ → 104 MeV while improvement is claimed), which is real and I agree it justifies CONDITIONAL. However, I judge the more load-bearing soft spot for the strongest_claim to be the normalization prescription in Eq. (16): the reader's concern questions whether the mixed model is a fair baseline, while mine questions whether the reported numbers are well-defined at all, independent of fitting philosophy. The two concerns are related (both stem from the Sec. II C decision not to redefine parameters) but distinct: even with a refit, the residual profile would still depend on which coupling is held fixed. I do not recommend moving the verdict to REJECT because (i) the mixing effect is a genuine, internally consistent feature of the model's gap/BSE system, (ii) the qualitative scaling with quark effective masses and their differences follows structurally from Eqs. (4)–(14) and is unlikely to be a normalization artifact, and (iii) the paper is transparent about the convention (\"Other normalization prescriptions can be explored in future studies\"). But CONDITIONAL is the right call, and the condition should explicitly include a prescription-robustness check in addition to the reader's refit demand. The concrete test above is cheap (a rerun of an existing numerical pipeline with one changed equation) and would settle whether the abstract's specific quantitative claims survive.","tokens_in":19556,"tokens_out":1685,"duration_ms":166505,"concrete_test":"Repeat the full pipeline (gap equations → BSE → Tables I–II → PDFs → NLO-DGLAP evolution) with one alternative normalization in Eq. (16): fix G44 = G0 (kaon reference) instead of G11, and optionally a symmetric prescription Gii = G0(G0+∆Gii)/[(G0+∆G11)(G0+∆G44)]^{1/2}. Then recompute the pion and kaon residual curves of Figs. 1–5. If the peaks at x≈0.3/0.8 (pion) and x≈0.2/0.7 (kaon) survive with the same sign and within ~50% in magnitude, the quantitative claim is prescription-robust; if the residuals relocate, shrink, or flip sign, the abstract's specific x-locations and magnitudes must be reported as convention-dependent.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central quantitative claim — mixing residuals of specific sign and size (~±1.5×10⁻³) peaking at x≈0.3 and 0.8 (pion) and x≈0.2 and 0.7 (kaon) — is generated entirely after imposing the normalization condition in Eq. (16), which fixes the charged-pion coupling G11 to the unmixed value G0 and pushes all mixing strength into the other diagonal couplings (G44, Guu, Gss). The text states explicitly: \"We fix the normalization by choosing the charged-pion coupling G11 as the reference value... Other normalization prescriptions can be explored in future studies.\" This is not a cosmetic choice. Because ∆Gij ∝ G0² times flavor-asymmetric loop integrals (Eqs. 4–14), the ratio (G0+∆Gii)/(G0+∆G11) that defines each normalized coupling changes non-trivially if a different reference channel is chosen (e.g., the kaon channel G44, or a symmetric average). Under a different prescription, the relative shifts of Mu, Ms, mπ, mK, and the meson-quark couplings gπqq, gKqq (Tables I–II) redistribute, and the PDF residuals — which are differences of O(10⁻³) between two full calculations — could shift in magnitude, move in x, or partially cancel. Since the residuals are themselves tiny compared to both the PDF values and the experimental uncertainties, even a modest prescription-induced reshuffling could erase or relocate the advertised peaks. This compounds the reader's flagged issue (no refit, mπ driven to 104 MeV): the model is simultaneously unrefitted and normalized by convention, so the claim that mixing \"certainly improves\" the PDFs rests on a comparison whose baseline and whose mixing pattern are both convention-laden. The existence of a nonzero mixing effect within the model is not in doubt; its quantitative profile — the actual content of the abstract — is.","agreement_with_reader":"partial"},"referee_report":{"model":"moonshotai/kimi-k3","summary":"The authors compute charged pion and kaon valence PDFs in the U(3) NJL model with proper-time regularization, adding flavor-dependent four-fermion couplings generated by vacuum polarization (Eqs. 4–15) rather than the 't Hooft determinant. The modified couplings enter the gap equations and Bethe–Salpeter equations, producing shifted constituent masses, meson masses, and meson–quark couplings (Tables I–II), which feed the standard NJL PDF calculation of Ref. [3] followed by NLO-DGLAP evolution to μ² = 4 and 27 GeV². The paper reports that mixing-induced residuals in the valence distributions are O(10⁻³), peaking near x ≃ 0.3 and 0.8 for the pion and x ≃ 0.2–0.3 and 0.7 for the kaon, and asserts that the mixing \"certainly improves\" the PDFs relative to the unmixed baseline when compared to E615 data and the JAM analysis.","tokens_in":20030,"tokens_out":2740,"duration_ms":113480,"significance":"If the results hold, the paper provides a controlled estimate of a genuinely parameter-conservative mixing mechanism: the vacuum-polarization-induced couplings ΔGij introduce no new energy scales or free parameters beyond those of the standard NJL model, and the work correctly preserves the valence and momentum sum rules (Eqs. 33–34) through the full pipeline. The predicted residual sizes and x-locations are, in principle, falsifiable with forthcoming EIC/EicC/J-PARC/AMBER data, and the comparison to both experimental data and the JAM global analysis is the right external benchmark. The gap/BSE + proper-time + NLO-DGLAP machinery is standard and internally consistent. The significance is tempered, however, by the fact that the advertised effects are ~10⁻³ in xq(x) — far below current experimental and global-fit uncertainties — so the practical impact rests entirely on the robustness of the residual calculation, which is where the manuscript is currently weakest.","major_comments":[{"comment":"The parameters (G0, ΛUV, ΛIR, current masses) are fitted to physical mπ and mK in the unmixed model and then held fixed when the mixing is switched on. As a result the pion mass collapses from 140 to 104 MeV (G19-Set4) and 121 MeV (G10-Set4), and mK shifts by ~15 MeV. The PDFs in Eq. (29) depend explicitly on m²ps and on the constituent masses that are themselves shifted; a comparison to data and JAM performed with a 104 MeV pion cannot support the claim (abstract and §IV) that mixing 'certainly improves' the PDFs. The improvement claim is load-bearing for the paper's conclusion. The authors should either (a) refit at least one parameter set with mixing active so that mπ and mK remain physical, and show whether the residuals and the improvement survive, or (b) remove the improvement claim and present the residuals strictly as a sensitivity estimate at fixed parameters.","section":"§II.C and Table I (G19-Set4, G10-Set4 rows)"},{"comment":"All quantitative results (residual magnitudes, signs, and peak locations) are generated after imposing the normalization G11 = G0, which the text itself describes as one choice among several ('Other normalization prescriptions can be explored in future studies'). Because the normalized couplings are ratios (G0+ΔGii)/(G0+ΔG11), choosing a different reference channel (e.g., G44, or a flavor-symmetric average) redistributes the shifts among Mu, Ms, mπ, mK, and the meson–quark couplings, and hence among the PDF residuals, which are differences of O(10⁻³) between two full calculations. Given that the residuals are tiny compared to both the PDF values and the data uncertainties, even a modest prescription-induced reshuffling could move or partially cancel the advertised peaks at x ≃ 0.3/0.8 (pion) and 0.2/0.7 (kaon). A robustness check — recomputing Tables I–II and one representative residual","section":"§II, Eq. (16)"},{"comment":"The statement that mixing 'certainly improves' the pion and kaon PDFs is not supported by any quantitative criterion in the manuscript. No χ², no comparison against the JAM uncertainty band, and no residual-with-respect-to-data is shown; the figures plot only the difference between the two model variants. With residuals of O(10⁻³) and data/JAM uncertainties orders of magnitude larger, an improvement at this level is indistinguishable from noise by any standard metric. The authors should either provide a quantitative goodness-of-fit comparison demonstrating the improvement or soften the claim to a statement about the size and location of the mixing effect.","section":"§IV (discussion of Figs. 1–2) and abstract"}],"minor_comments":[{"comment":"The abstract states the dominant kaon mixing effects appear at x ≃ 0.2 and x ≃ 0.7, but the body text (discussion of Figs. 4–6) repeatedly identifies the dominant kaon residual at x ≃ 0.3 and describes the x ≃ 0.8 region as 'significantly suppressed'. The abstract and body should be reconciled.","section":"Abstract vs. §IV"},{"comment":"The text reads 'we show the difference of the kaon up valence quark distribution at scale μ² = 27 GeV² for fixed G0 = 10 GeV⁻²' while referring to Fig. 6(b), which is labeled μ² = 4 GeV². Please correct the scale reference.","section":"§IV, paragraph on Fig. 6(b)"},{"comment":"The statement that the kaon PDF difference is 'one order of magnitude larger' than the pion's is not borne out by the figure axes: pion residuals peak at ~±0.0015–0.002 (Figs. 1, 3) and kaon residuals at ~±0.006 (Fig. 6), i.e., a factor of roughly 3–4. Please correct.","section":"§IV, kaon vs. pion residuals"},{"comment":"The coupling is written as 'G0 = 10 GeV²' in several places; the correct unit is GeV⁻² (as used elsewhere). Please make the units uniform.","section":"§IV, multiple occurrences"},{"comment":"The NLO running-coupling expression is written in a nonstandard compact notation (β3, ln ln(QΛ)/ln(QΛ)); please check for typos and consider giving the standard two-loop form with β1/(4πβ0) coefficient for clarity.","section":"§III, Eq. (39)"},{"comment":"The residual panels would be more informative with the JAM uncertainty band (or E615 error bars) overlaid, so the reader can directly see that the mixing residuals are below current sensitivity; this would also discipline the 'improvement' language.","section":"Figs. 1–9"},{"comment":"In the isospin-symmetric limit used here, Eqs. (18) and (19) are identical; it would help the reader to state explicitly that Guu = Gdd by construction and that all u/d asymmetry in the PDFs comes from the meson–quark vertex and mass terms, not from the couplings.","section":"§II, Eqs. (18)–(20)"}],"recommendation":"major_revision","confidential_remarks":"The calculation pipeline is competent and the paper sits naturally in the journal's scope, but the two central quantitative selling points — the 'certain improvement' over the baseline and the specific residual peak locations — currently rest on an unrefitted parameter set (mπ = 104 MeV in the flagship mixed set) and an arbitrarily chosen coupling normalization. Both are fixable within the manuscript's existing framework, hence major revision rather than rejection. I also note a heavy reliance on the first author's prior work for the mixing formalism (Refs. 34–39); this is not improper, but the novelty over those papers is essentially the PDF application, and the editor may wish to weigh that when assessing fit."},"author_rebuttal":null,"desk_editor":{"model":"grok-4.5","letter":"The new piece here is straightforward: they take the vacuum-polarization flavor-mixing couplings from their earlier NJL papers and run them through charged pion and kaon valence PDFs, then evolve to 4 and 27 GeV² and difference against the unmixed baseline. That joint calculation had not been done. The pipeline (gap + BSE, proper-time, NLO-DGLAP) is standard and internally consistent; sum rules are kept, residual plots are readable, and the comparison to Conway and JAM is honest.\n\nWhat they do well is isolate a small, x-dependent shift whose size tracks the effective-mass differences, with peaks near the x values advertised in the abstract. Inside the model the effect is real and of the same order as some CSV studies they cite. For people who already work in meson PDFs or who need a concrete systematic for EIC/EicC/AMBER, that is usable information.\n\nTwo soft spots matter, in proportion. First, they explicitly do not refit after mixing is turned on. Table I shows mπ collapsing from 140 to 104 MeV (G19-Set4) while the text still says the mixing “certainly improves” the PDFs. That language is stronger than the evidence. Second, the quantitative residual profile sits on the normalization choice in Eq. (16) that freezes G11 = G0 and pushes mixing into the other diagonals; they flag that other prescriptions are possible. Because the residuals are O(10⁻³), a different reference channel could move or shrink the advertised peaks. Neither issue kills the existence of a mixing effect; both weaken the claim that the specific shape and the improvement are robust.\n\nNo code, no uncertainty bands on the residuals, self-citation of the mixing kernels is appropriate given prior work. This is for the effective-model meson-structure crowd, not for global PDF fits. It deserves a serious referee who will demand a refit (or a clear statement that the mass shift is left unrepaired) and a check under at least one alternate normalization. I would send it to review, expect revision on the wording, and keep the residual plots in the literature once cleaned up.","headline":"Real but tiny mixing residuals on pion/kaon PDFs inside NJL; useful extension, undercut by no refit (mπ→104 MeV) and a flagged normalization convention.","tokens_in":20701,"tokens_out":578,"would_cite":false,"duration_ms":24171,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"grok-4.5","headline":"Vacuum-polarization flavor mixing shifts charged pion and kaon valence PDFs in proportion to quark mass differences and improves the distributions at both 4 and 27 GeV².","keywords":["parton distribution functions","pion","kaon","Nambu–Jona-Lasinio model","flavor mixing","vacuum polarization","DGLAP evolution","charge symmetry"],"falsifier":"A high-precision measurement of the charged-pion or kaon valence PDF near x ≈ 0.3 and x ≈ 0.8 (or 0.2 and 0.7 for the kaon) at a few-GeV scale that shows no residual of the size and shape predicted by the mixed versus unmixed NJL curves would falsify the claimed improvement.","tokens_in":20351,"feed_emoji":"⚛️","tokens_out":955,"duration_ms":17328,"temperature":0.7,"pith_summary":"This paper asks how a flavor-mixing interaction generated by vacuum polarization, rather than by instantons, changes the valence-quark parton distributions of the charged pion and kaon. Working inside the U(3) Nambu–Jona-Lasinio model with proper-time regularization, the authors recompute dynamical quark masses, meson masses and meson–quark couplings once the mixing is turned on, then evolve the resulting valence PDFs to the experimental scales 4 and 27 GeV². They find that the size of the mixing correction tracks the effective quark masses and their differences, producing the largest residuals near x ≈ 0.3 and 0.8 for the pion and near x ≈ 0.2 and 0.7 for the kaon. The same pattern appears at both scales, and the mixed distributions sit closer to existing data and global analyses than the unmixed baseline. A sympathetic reader cares because forthcoming electron–ion collider measurements will finally have the precision to test whether such vacuum-induced mixing is visible in light-meson structure.","feed_headline":"Flavor mixing shifts pion and kaon PDFs at x~0.3 and 0.8","feed_subtitle":"Vacuum-polarization effects track quark mass differences and improve agreement with data at 4 and 27 GeV²","key_machinery":"Implicit flavor mixing: diagonal couplings G_ff acquire dependence on the constituent masses of the other flavors through vacuum-polarization loops, normalized so that the charged-pion coupling remains the reference value G_11 = G_0.","core_discovery":"The strength of vacuum-polarization flavor-mixing effects on charged-pion and charged-kaon valence PDFs is proportional to the quark effective masses and their differences; the dominant residuals lie near x ≃ 0.3 and 0.8 (pion) and x ≃ 0.2 and 0.7 (kaon) at both μ² = 4 and 27 GeV², and the mixing improves the PDFs relative to the unmixed NJL baseline.","pith_inferences":["Because the residual strength is essentially scale-independent between 4 and 27 GeV², the mixing signature is already fixed at the model scale and is only mildly diluted by DGLAP evolution.","Refitting the ultraviolet cutoff and bare couplings after mixing is turned on would likely restore the physical pion mass while preserving the shape of the PDF residuals, offering a clean next calculation.","The same implicit-mixing mechanism should generate analogous shifts in the pion and kaon electromagnetic form factors and generalized parton distributions."],"forward_implications":["Vacuum-polarization mixing, not only the ’t Hooft determinant, must be included when extracting light-meson PDFs from effective models.","The same mass-difference scaling implies larger mixing residuals for heavier flavor combinations once data become available.","Lattice calculations of pion and kaon PDFs can isolate the quark-mass-difference piece of flavor mixing by comparing isospin-symmetric and broken ensembles.","Upcoming EIC, EicC, J-PARC and AMBER Sullivan/Drell–Yan data will be able to test the predicted residual shapes directly."],"fun_headline_variants":["Flavor mixing shifts pion kaon PDFs near x=0.3 and 0.8","Vacuum-polarization mixing tracks quark mass gaps in pion kaon PDFs","Mixing effects peak at x~0.3/0.8 pion and 0.2/0.7 kaon","NJL flavor mixing improves pion kaon valence PDFs at 4 and 27 GeV²","Quark-mass differences set mixing strength in charged meson PDFs"],"cache_read_input_tokens":16512,"weakest_assumption_plain":"The authors keep the original model parameters fixed and simply insert the new flavor-dependent couplings, even though this shifts the pion mass well below its physical value.","fun_headline_variants_meta":{"raw":{"variants":["Flavor mixing shifts pion kaon PDFs near x=0.3 and 0.8","Vacuum-polarization mixing tracks quark mass gaps in pion kaon PDFs","Mixing effects peak at x~0.3/0.8 pion and 0.2/0.7 kaon","NJL flavor mixing improves pion kaon valence PDFs at 4 and 27 GeV²","Quark-mass differences set mixing strength in charged meson PDFs"]},"model":"grok-4.5","effort":"low","cost_usd":0.00379,"raw_usage":{"total_tokens":1252,"prompt_tokens":876,"num_sources_used":0,"completion_tokens":99,"cost_in_usd_ticks":37904000,"prompt_tokens_details":{"text_tokens":876,"audio_tokens":0,"image_tokens":0,"cached_tokens":128},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":277,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":876,"tokens_out":99,"duration_ms":5486,"temperature":1.0,"reasoning_tokens":277,"cache_read_input_tokens":128,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-30T20:43:39.943217+00:00","model_set":{"reader":"grok-4.5"},"falsifier":"A high-precision measurement of the charged-pion or kaon valence PDF near x ≈ 0.3 and x ≈ 0.8 (or 0.2 and 0.7 for the kaon) at a few-GeV scale that shows no residual of the size and shape predicted by the mixed versus unmixed NJL curves would falsify the claimed improvement.","supporting_citations":[],"review_version":1}