{"id":"861fd16a-f68f-419a-b0e6-3f9c04ec3d9b","arxiv_id":"2602.14747","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":8,"one_line_summary":"A Bayesian Jeans analysis of three ultra-diffuse galaxies finds no preference for a non-minimal dark-matter-gravity coupling and yields weak upper limits on its length scale.","lead":"The paper tests a modified-gravity idea in which dark matter couples to spacetime curvature, using the internal motions of three ultra-diffuse galaxies. It finds no evidence for the coupling and reports only upper limits, meaning current data cannot yet distinguish this model from standard gravity.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The Bayesian upper limits on L are not interpretable because the prior on L—the key new parameter—is never reported in Table II or the text; the central quantitative claim is therefore prior-dependent.","rationale":"The reader's declared weakest assumption is the equilibrium/spherical Jeans assumption, but their rationale also lists the unreported L prior and boundary-hitting upper limits. My stress test identifies the missing L prior as the most load-bearing concern because it directly undermines the quantitative content of the central claim—the upper limits—rather than only the physical interpretation. The equilibrium assumption is standard in Jeans analyses and is a known limitation, though it is not internal to the paper and is harder to test with existing data. In contrast, the prior on L is completely absent from Table II despite being the key new parameter. Without it, the reported numbers are not reproducible, and the statement that 'L shows no preference for non-zero values' could be an artifact of a prior concentrated near zero. This is a fixable reporting/robustness issue rather than a fundamental flaw, so the conditional verdict stands unchanged.","tokens_in":26962,"tokens_out":4688,"duration_ms":53753,"concrete_test":"Obtain/report the L prior and rerun one representative case (e.g., DF44 gNFW + SHMR + constant anisotropy, ϵ=-1) with at least two priors: (i) log-uniform over a wide range, e.g., log10(L/kpc) ∈ [-5,5]; (ii) uniform in L over [0,100] kpc. If the 95% upper limit on L changes by more than a factor of ~2, or if the posterior accumulates at the prior boundary, the quoted upper limits are prior-dominated and the 'sensitivity limit' interpretation is not supported. Also verify that the GR control case reproduces the same astrophysical parameters under both priors.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central quantitative claim is that the NMC length scale L is consistent with zero and can only be bounded from above. However, Table II, which is intended to list all priors, contains no prior for L, and no L prior is stated anywhere in the text or appendices. In a Bayesian analysis, the posterior of L and the evidence ratios are conditional on this prior. Without knowing whether L was assigned a uniform prior in linear space, a log-uniform prior, a normal prior, or a bounded interval, the reported upper limits (e.g., logL < 0.30 for DF2; logL < -5.14 or < -109.05 for DF4; logL < -2.08 for DF44) cannot be reproduced, compared with the Gandolfi et al. [45] relation, or interpreted as a physical 'sensitivity limit.' Several reported values are extreme negative numbers or 'U' flags, suggesting prior-boundary behavior or poorly constrained directions; without the prior these cannot be distinguished from a genuine constraint. The broad no-detection conclusion is likely robust—GR and NMC fits are statistically indistinguishable—but the quantitative upper limits, which are the paper's headline output, are not self-contained as presented.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper tests a non-minimal coupling (NMC) between dark matter and curvature using spherical Jeans models of three ultra-diffuse galaxies: NGC 1052-DF2, NGC 1052-DF4, and Dragonfly 44. The NMC correction is absorbed into an effective mass profile (Eq. 3.9) proportional to L^2 ∇^2 ρ_DM. For each galaxy the authors run Bayesian MCMC/Nested Sampling over multiple halo profiles, two anisotropy models, with and without a stellar-to-halo mass relation, and the two polarization signs ϵ=±1. The principal result is that GR and NMC fits are statistically indistinguishable, the L posterior is always compatible with zero and yields only upper limits, and the inferred astrophysical parameters agree with GR literature values. A mock DF44 sensitivity test indicates that only large L values produce a detectable signal. The paper concludes that current UDG kinematics provide only weak sensitivity limits on the NMC length scale.","tokens_in":27346,"tokens_out":9763,"duration_ms":98937,"significance":"If the result holds, it is a useful null result: it extends NMC constraints from galaxy clusters to individual galaxies and demonstrates that UDGs, despite their extreme dynamical environments, cannot yet discriminate this coupling. The study is thorough in its exploration of halo profiles, anisotropy prescriptions, and SHMR variants; the effective-mass formulation is clean; the GR baselines are checked against literature values; and a mock-data sensitivity test is included. However, the headline quantitative upper limits on L cannot be reproduced or interpreted without the prior on L, and several reported limits appear to sit at prior boundaries. The no-detection conclusion is likely robust, but the quantitative sensitivity claims are not self-contained as presented.","major_comments":[{"comment":"The prior on the new parameter L is never reported. Table II lists priors for c200, M200, γ, β, r_a, M*, D, v_sys, and Υ*, but not for L, and no statement is given in the text about whether L is uniform in linear or log space, its bounds, or its sign. Because the headline output is a set of Bayesian upper limits (Tables III–IX), and because those limits depend on the prior volume, the reported numbers are not reproducible. Values such as logL < −109, < −38, and < −33 (Tables IV, VII) cannot be distinguished from prior-boundary artifacts without knowing the prior. Please state the L prior explicitly and test sensitivity to its width and type.","section":"Table II / Sec. V"},{"comment":"The prior for the anisotropy radius r_a is listed as U(0,∞). As written, this is an improper prior, so the Bayesian evidences used for model comparison (logB in Tables III–IX) are not strictly well-defined. If a large finite upper bound is used in practice, it must be reported; otherwise the evidence ratios should be treated as provisional. The same concern applies to any unbounded direction in the L prior.","section":"Table II"},{"comment":"The printed gNFW enclosed mass appears to omit the 1/(3−γ) prefactor that arises from ∫_0^x t^{2−γ}(1+t)^{γ−3} dt, and the normalization ρ_s is typeset in a garbled way with a hypergeometric argument (γ−2)c_Δ. If the code implements the standard formula, this is a typo, but as written the equation is not reproducible and is central to every halo fit. Please correct the formula and verify that the code matches the standard expression.","section":"Eq. (3.14)"},{"comment":"The claim of a universal, small L is stronger than the presented constraints. The largest-L selections in Table IX include logL = 22.15 for the GPI profile (DF44), and cNFW yields unconstrained L; within the main tables, upper limits vary by many orders of magnitude across galaxies and configurations (e.g., Table III logL<0.30 vs. Table IV logL<−109). The conclusion should be rephrased as a no-detection statement with configuration-dependent upper limits, rather than a universal small-L result.","section":"Table IX / Sec. VI"},{"comment":"The paper acknowledges that without the SHMR prior, DF2 and DF4 develop 'astrophysically implausible' halos (logM200 ∼ 4–5, c200 ∼ 23–25). Since the NoSHMR runs are used to argue that the near-GR outcome is not an artifact of halo modeling, and since Table IX selects high-L cases from these runs, the authors should demonstrate that the L upper limits are not driven by the unphysical parameter region—for example, by repeating the analysis with a physically motivated lower bound on M200 or by showing that the high-L posterior mass lies in the plausible region.","section":"Sec. VI / Tables III–IV"}],"minor_comments":[{"comment":"The analysis assumes spherical, non-rotating equilibrium and treats the globular clusters and DF44 radial bins as relaxed tracers. The paper does not test for interloper contamination or flattening; this should be acknowledged as a limitation affecting the absolute scale of the L limits.","section":"Sec. III / Sec. VI"},{"comment":"The mock DF44 test is generated with the same model and noise statistics as the real observations, so it calibrates ideal sensitivity but not robustness to model misspecification. This should be stated explicitly in the text.","section":"Sec. VI"},{"comment":"The heading 'NGC1052-DF44' should read 'Dragonfly 44' to match the rest of the paper.","section":"Table V heading"},{"comment":"The statement 'the polarization parameter ϵ likewise has little impact, as the cases ϵ=±1 yield comparable constraints on L' is difficult to reconcile with Table IV, where logL ranges from <−5.14 to <−109 between the two polarizations. Please clarify whether this refers to the qualitative conclusion rather than the numerical limits.","section":"Sec. VI"},{"comment":"The text says 'we do not report individual evidence values in the tables', yet Tables III–IX contain logB columns. Please rephrase to avoid the apparent contradiction.","section":"Sec. VI"}],"recommendation":"major_revision","confidential_remarks":"The missing prior on L is the main obstacle. If the authors supply the prior, show that the extreme upper limits are not prior-boundary artifacts, and correct the gNFW formula, I would support publication. The no-detection conclusion is credible, but the quantitative headline is not currently reproducible."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a competent null-result paper, and the broad conclusion—no detection, GR and NMC statistically indistinguishable—is probably right. But the headline quantitative output, the upper limits on L, is not self-contained. The stress-test note is correct: Table II lists the priors and there is no prior for L anywhere in the text or appendices. That matters, because posterior upper limits and evidence ratios inherit the prior. Without knowing the prior range and scale, the reported values like logL < -109 or < -33 cannot be reproduced or compared with the Gandolfi relation. Some of those extreme numbers look like prior-boundary or convergence artifacts; without the prior you cannot tell.\n\nWhat is genuinely new: this is the first application of the Bettoni-Liberati NMC framework to UDG kinematics, using three galaxies that span dark-matter-poor to dark-matter-dominated. The analysis covers multiple halo profiles, two anisotropy models, SHMR on/off, and includes a mock DF44 sensitivity test. The effective-mass treatment in Eq. (3.9) is internally consistent, and the GR baselines match published values. The mock test is useful: it shows current kinematics are only sensitive to large couplings, which are already disfavored. The authors are also appropriately cautious—they frame the upper limits as sensitivity bounds, not exclusions, and they openly flag the Bayesian-evidence inconsistency in DF44.\n\nSoft spots, in proportion: first, the missing L prior—this is the main issue and needs fixing before the quantitative bounds can be trusted. Second, the tracer samples are tiny (10 and 7 globular clusters, 9 radial bins) and the equilibrium spherical assumption is not tested; that is a structural limitation, though the authors acknowledge the sparse data. Third, some L upper limits are extreme negative numbers or 'U' flags, suggesting degeneracy or prior-boundary behavior; these should not be presented as constraints without diagnostics. Fourth, no code or data release, which is a shame for reproducibility.\n\nThe no-detection conclusion is likely robust—astrophysical parameters stay consistent with GR across configurations. But the L bounds need prior specification and robustness checks before being used as reference values. This paper deserves a serious referee; I would send it to review with requests to add the L prior, test prior sensitivity, and show corner plots or traces for the problematic runs. Worth engaging, but treat the L limits as provisional.","headline":"Solid null-result paper on NMC from UDG kinematics, but the central L upper limits are not interpretable because the prior on L is never reported.","tokens_in":27790,"tokens_out":2224,"would_cite":false,"duration_ms":27292,"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":"This paper argues that current kinematic data for ultra-diffuse galaxies cannot distinguish a non-minimal dark-matter–gravity coupling from General Relativity, yielding only upper limits on the coupling length scale.","keywords":["non-minimal coupling","dark matter","ultra-diffuse galaxies","Jeans modeling","modified gravity","velocity dispersion","Dragonfly 44"],"falsifier":"Re-analyze the same three galaxies with a model that relaxes spherical symmetry (e.g., an axisymmetric potential) or removes interloper contamination; if the posterior for L then excludes zero, the paper's null conclusion would be falsified.","tokens_in":26854,"feed_emoji":"🌌","tokens_out":10374,"duration_ms":97335,"temperature":0.7,"pith_summary":"The paper asks whether dark matter's non-minimal coupling to spacetime curvature—an extension of General Relativity with a coupling length scale L—can be detected in the internal motions of ultra-diffuse galaxies, the most diffuse galaxies known. It models three UDGs spanning the range from dark-matter-deficient (NGC 1052-DF2, NGC 1052-DF4) to dark-matter-dominated (Dragonfly 44), using Bayesian Jeans analysis with eight dark-matter halo profiles, two orbital anisotropy models, and with and without stellar-to-halo mass priors. The result is a null detection: across all configurations, the inferred astrophysical parameters match their General Relativity counterparts, and the posterior for L is always compatible with zero, giving only upper limits. The authors stress these limits are a sensitivity limit of current data, not a tight exclusion, because UDGs have shallow density profiles that suppress the coupling correction, which scales with the Laplacian of the dark-matter density.","feed_headline":"No sign of non-minimal dark-matter coupling in faint galaxies","feed_subtitle":"Three extreme galaxies - from dark-matter-free to dark-matter-dominated - all fit General Relativity; the coupling scale stays an upper limi","key_machinery":"The central object is the effective mass profile M_eff(r)=M(r)−4π ε r² L² dρ_DM/dr, derived from the modified Poisson equation ∇²Φ=4πG[ρ_tot − ε L² ∇²ρ_DM]. This lets the standard spherical Jeans equation be applied with M_eff in place of M, converting the non-minimal coupling into an effective force-law change. The correction scales as L² times the dark-matter density's Laplacian; because ultra-diffuse galaxies have shallow density profiles over the observed radii, the term is tiny, which is why the data cannot distinguish the non-minimal coupling from General Relativity.","core_discovery":"Three ultra-diffuse galaxies—NGC 1052-DF2, NGC 1052-DF4, and Dragonfly 44—show no statistical preference for a non-minimal dark-matter–gravity coupling. The model changes the Poisson equation by a term proportional to L²∇²ρ_DM, equivalent to an effective mass M_eff(r)=M(r)−4π ε r² L² dρ_DM/dr. Using this in the Jeans equation, the analysis recovers General Relativity parameters across eight halo profiles and two anisotropy models. The coupling length L yields only upper limits, which the authors interpret as a sensitivity limit of the sparse globular-cluster and stellar velocity data rather than as a tight exclusion of the coupling.","pith_inferences":["A logical extension the authors do not develop: if L is genuinely small, the most promising observational targets for NMC are systems with sharp dark-matter density gradients (e.g., cuspy dwarf spheroidals or cluster cores), where the Laplacian of the density is large, rather than the diffuse regions where UDGs live.","The mock-data sensitivity test implies that merely adding more tracers to the same UDGs may not break the degeneracy; gains in velocity precision per tracer, or selecting galaxies with steeper inner profiles, would be more effective.","The spread in Dragonfly 44 Bayesian evidences (e.g., NMC logB ≈ 0.64–0.65 vs GR −0.31 despite similar posteriors) suggests the evidence estimates carry systematic uncertainty; repeating the comparison with an independent nested-sampling implementation or an information criterion would clarify whether any model is actually favored."],"forward_implications":["If the non-minimal coupling exists, its characteristic scale on galactic scales is below the current detection threshold; only upper limits can be set, not a measurement.","The GR and NMC fits are statistically indistinguishable, so the mass–anisotropy degeneracy and SHMR-prior effects seen in DF2 and DF4 (extremely low halo mass with high concentration) are unaffected by the coupling.","The L upper limits are similar across dark-matter-poor and dark-matter-dominated systems, hinting at a nearly universal coupling scale if combined with cluster-scale results, though the sample is small.","Future high-precision velocity measurements—more globular clusters or finer radial bins—are required to determine whether non-minimal coupling effects can be distinguished in low-acceleration systems."],"fun_headline_variants":["No dark-matter gravity coupling seen in three ultra-diffuse galaxies","Ultra-diffuse galaxies stick to general relativity, no modified coupling","Dark-matter–curvature coupling absent in faintest galaxies","Non-minimal coupling? Extreme galaxies show none, GR wins","Three ultra-diffuse galaxies find no trace of modified gravity"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The analysis assumes the globular-cluster and stellar velocities in the three galaxies are equilibrium tracers of a spherical, non-rotating potential; if the galaxies are flattened, unvirialized, or contaminated by interlopers, the inferred velocity dispersions—and therefore the L upper limits—could be biased.","fun_headline_variants_meta":{"raw":{"variants":["No dark-matter gravity coupling seen in three ultra-diffuse galaxies","Ultra-diffuse galaxies stick to general relativity, no modified coupling","Dark-matter–curvature coupling absent in faintest galaxies","Non-minimal coupling? Extreme galaxies show none, GR wins","Three ultra-diffuse galaxies find no trace of modified gravity"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000214,"raw_usage":{"total_tokens":1298,"prompt_tokens":820,"completion_tokens":478,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":564,"completion_tokens_details":{"reasoning_tokens":387}},"tokens_in":564,"tokens_out":478,"duration_ms":5519,"temperature":1.0,"reasoning_tokens":387,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-04T05:59:11.503221+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Re-analyze the same three galaxies with a model that relaxes spherical symmetry (e.g., an axisymmetric potential) or removes interloper contamination; if the posterior for L then excludes zero, the paper's null conclusion would be falsified.","supporting_citations":[],"review_version":1}