{"id":"64ae3262-f722-4d1b-b358-b86c8001f589","arxiv_id":"2510.16228","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Late-time data bound the T-duality zero-point-length coupling to β ≲ 10^-3, leaving ΛCDM statistically equivalent.","lead":"This paper constrains a string-theory-inspired tweak to the Friedmann equations, parameterized by β, using supernovae, BAO, cosmic chronometers, and gamma-ray bursts. It finds β ≲ 10^-3 and that the modified model fits the data as well as standard ΛCDM.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The reported β upper bound is ambiguous because the paper never specifies whether the likelihood uses the exact solution (13) or the expansion (14); at z=8.1 and β≈3×10⁻³ the exact model is imaginary, so the high-z GRB constraint is not well defined.","rationale":"The reader's weakest-assumption diagnosis is exactly the load-bearing issue: the paper's headline bound depends on how the likelihood handles the square root in Eq. (13) at high redshift. This is not a mere technicality. The GRB sample extends to z=8.1, where the exact model ceases to exist for β values within or above the reported 68% interval. If the exact solution and a hard cutoff were used, the posterior would be truncated at β≈1.1×10⁻³, tightening the bound but for a different reason than distance-modulus fitting. If the expansion were used, the likelihood would be evaluating a first-order formula in a regime where βD is not small, so the model is not the one presented. In either case, the statistical comparison with ΛCDM for D3/D6 is not well defined as written. This does not necessarily overturn the qualitative conclusion that β is small, but it does undermine the specific numerical upper bounds and the AIC comparison involving GRBs. The paper is otherwise a standard, clearly written parameter-constraint study, and no code or chains are released, which makes the missing likelihood specification harder to resolve by inspection. The conditional verdict already captures the need for clarification; no further adjustment is needed.","tokens_in":11796,"tokens_out":4191,"duration_ms":39566,"concrete_test":"Re-run the D3 chain (PP+OHD+BAO+GRB) with two explicit likelihood implementations: (a) exact Eq. (13) with zero likelihood whenever 1−4βD(z)<0; (b) first-order expansion Eq. (14) restricted to redshifts where βD(z)<0.1. Report the 68% upper bounds on β and compare with the published <0.0034. Also compute z_max(β=0.003, Ωm0=0.30, ΩΛ0=0.70) from 4βD(z)=1; if z_max<8.1, the z≈8.1 GRB lies outside the exact model's domain. If the two implementations give meaningfully different bounds, or if the exact-solution chain excludes β>≈10⁻³ because of the high-z GRBs, the published bound is an artifact of an unspecified likelihood choice.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central quantitative result is the 68% upper bound β≲3–4×10⁻³ for the datasets containing GRBs (D3 and D6 in Table II). This bound is not well defined until the likelihood is specified. The exact solution Eq. (13) exists only when 1−4βD(z) ≥ 0, with D(z)=Ωm0(1+z)³+Ωr0(1+z)⁴+ΩΛ0. Using the reported best-fit Ωm0≈0.30 and ΩΛ0≈0.70, at z=8.1 one has D≈227, so even β=0.003 gives 4βD≈2.7>1 and the square root in Eq. (13) is imaginary. Thus, for β near the reported upper bound, the exact model has no real H(z) for the highest-redshift GRBs (z≈8.1) in the Amati-calibrated sample [71]. The paper never states whether the MCMC likelihood uses the exact expression (13) or the first-order expansion (14), nor how the imaginary regime is handled. If Eq. (13) is used with a hard existence cutoff, the GRB likelihood excludes β≳1/(4D_max)≈1.1×10⁻³, which is inconsistent with the quoted <3.4×10⁻³. If Eq. (14) is used instead, it is evaluated where βD≈0.7, far outside the O(β) validity of the expansion, so the high-z model is not the one being constrained. Either way, the inference drawn from the GRB-containing combinations is not well defined.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper derives a modified Friedmann equation from a T-duality-inspired correction to horizon entropy, introducing a dimensionless parameter β ∼ l0^2 H0^2. Using Cobaya with MCMC sampling, the authors constrain β with six combinations of late-time datasets: PantheonPlus or Union3 supernovae, cosmic chronometers, DESI DR2 BAO, and Amati-calibrated GRBs. The main result is an upper bound β ≲ O(10^-3) (68% C.L.) for combinations including BAO, with AIC showing statistical equivalence to ΛCDM. The paper does not state whether the likelihood uses the exact Hubble solution (13) or the first-order expansion (14).","tokens_in":12213,"tokens_out":6378,"duration_ms":51400,"significance":"If the reported constraints are correct, this constitutes the first quantitative late-time observational bound on string T-duality inspired cosmology, complementing earlier early-universe analyses in Ref. [37]. The use of multiple recent datasets, a standard Bayesian pipeline, and AIC model comparison are strengths. However, the central quantitative result is currently not well defined because the high-redshift behavior of the model depends on which of the two expressions (13)/(14) is implemented; for the GRB-containing datasets, the exact solution becomes imaginary for β near the reported upper bound. This issue must be resolved before the significance of the constraints can be assessed.","major_comments":[{"comment":"The paper never states whether the MCMC likelihood uses the exact Hubble solution (13) or the first-order expansion (14). This ambiguity is load-bearing for the central claim. For the GRB-containing combinations D3 and D6, the highest-redshift data point is z=8.1. Using the reported best-fit Ωm0≈0.30, ΩΛ0≈0.70 and a standard Ωr0≈9×10^-5, D(z)≡Ωm0(1+z)^3+Ωr0(1+z)^4+ΩΛ0 is ≈227 at z=8.1, so 4βD≈2.7 for β=0.003, making the square root in Eq. (13) imaginary. If Eq. (13) is used with a hard existence cutoff, the likelihood excludes β≳1/(4D_max)≈1.1×10^-3, which is inconsistent with the quoted <3.4×10^-3 in Table II. If Eq. (14) is used instead, it is evaluated at βD≈0.7, far outside the O(β) validity of the expansion. Either way, the reported upper bounds from the GRB datasets are not well defined. The authors must specify the likelihood implementation and, if Eq. (13) is used, the prior/cuto","section":"Sec. 2, Eq. (9) and Eq. (15)"},{"comment":"The derivation of Eq. (9) is summarized as 'straightforward algebra' without presenting intermediate steps. Since Eq. (9) is the foundation of all constraints, the reader cannot verify the approximations, e.g., the dropping of O(α^2) terms. In addition, the flatness condition is imposed differently in the exact and expanded formulations: imposing H(0)/H0=1 on Eq. (13) gives ΩΛ0=1−Ωm0−Ωr0−β exactly, whereas Eq. (15) contains O(β^2). The text should state which expression is used in the likelihood and whether this choice affects the reported parameter bounds.","section":"Sec. 2, Eq. (9) and Eq. (15)"}],"minor_comments":[{"comment":"The priors do not include the SNIa absolute magnitude or any GRB calibration nuisance parameter. Please clarify how the distance-modulus likelihood is normalized (e.g., analytic marginalization over M), especially since PantheonPlus and Union3 are used without SH0ES calibration.","section":"Sec. 3.2 / Table I"},{"comment":"The text introduces Ω_r0 but does not specify how it is fixed or sampled. Please state the value or prior used for the radiation density parameter.","section":"Sec. 2.1"},{"comment":"In the rows for D1 and D4, the '−' for r_drag is ambiguous; use 'N/A' or a dash to indicate that BAO data are not included.","section":"Table II"},{"comment":"The claim of 'first quantitative observational constraints' should be qualified as 'first late-time' since Ref. [37] already provides early-universe constraints on the same framework.","section":"Abstract"}],"recommendation":"major_revision","confidential_remarks":"The central issue is the missing specification of the likelihood implementation. The manuscript currently permits two different models (exact and expanded) that produce different predictions at high redshift; the reported GRB-containing bounds depend critically on this choice. I recommend requiring the authors to clarify and, ideally, release the likelihood code. The paper would also benefit from making the derivation of Eq. (9) self-contained enough for the reader to verify."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's the honest read.\n\nThis is a competent, fairly standard Bayesian constraints paper. What is new is not the theory—the modified Friedmann equation and entropy came from the authors' own earlier paper—but the first late-time observational bound on the T-duality length parameter. If you believe the theoretical setup, the number β≲O(10^-3) is a useful datum, and the overall conclusion that the model is statistically indistinguishable from ΛCDM is credible. The dataset choices are defensible, the inclusion of DESI DR2 is timely, and the AIC comparison is handled properly.\n\nThe main problem is a missing specification. Nowhere does the paper say whether the likelihood uses the exact Hubble solution (13) or the first-order expansion (14). This matters. At z=8.1, with the best-fit Ωm0≈0.30, D≈227. The exact solution is real only if 1−4βD≥0, i.e. β≲1.1×10^-3. The paper quotes β<3.4×10^-3 for the GRB-containing combination. So either the exact solution is evaluated outside its domain, or the expansion is evaluated at βD≈0.7, where first order is not a good approximation. The GRB results are therefore not well defined as written. The no-GRB combination (D2) gives the same order of bound and does not have this issue, so the central claim probably survives—but the GRB section needs to be redone or explicitly flagged as relying on the expansion with the domain restriction stated.\n\nMinor issues: the drop from the first law to Eq. (9) is just called straightforward algebra, no code or chains are released, and the 'first constraints' framing should be read narrowly—it is a standard MCMC exercise applied to a model the same group introduced. The self-citation is fine because that is the actual source of the model.\n\nWho should read this? People testing minimal-length/quantum-gravity-inspired cosmologies with late-time data. It will be citable as the current late-time bound on this specific model. It does not resolve any major problem and is not methodologically novel.\n\nI would send it to a serious referee. The ambiguity is fixable—state the likelihood, rerun with a domain-consistent cutoff, or justify the expansion—and the paper would be acceptable after that.","headline":"A competent constraint paper whose headline β bound is probably right, but the GRB fits have a model-domain ambiguity that needs fixing before publication.","tokens_in":12660,"tokens_out":3418,"would_cite":false,"duration_ms":33381,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["83F05","83C45","85A40"],"pacs":["98.80.-k","04.60.-m","95.36.+x"],"model":"deepseek-v4-flash","headline":"A string-theory-inspired correction to gravity is constrained to be tiny by late-time cosmological data.","keywords":["string T-duality","zero-point length","modified Friedmann equations","cosmological constraints","dark energy","Hubble parameter","baryon acoustic oscillations","gamma-ray bursts"],"falsifier":"Recompute the distance modulus of the highest-redshift GRB (z ≈ 8.1) using the exact solution (13) versus the first-order expansion (14) at β = 0.003; if the two predictions differ by more than the reported measurement uncertainty, the upper bound on β depends on which expression was used in the likelihood, and the constraint is not purely data-driven.","tokens_in":11624,"feed_emoji":"🔭","tokens_out":2475,"duration_ms":22741,"temperature":0.7,"pith_summary":"This paper works out the observable consequences of a cosmological model motivated by string T-duality, where a zero-point length l0 modifies the gravitational potential and hence the Friedmann equations of an expanding universe. The modification shows up as a single dimensionless parameter β, proportional to l0²H0², which measures how much the expansion history departs from the standard ΛCDM model. Combining supernova distances, cosmic chronometer Hubble measurements, baryon acoustic oscillations, and calibrated gamma-ray bursts, the authors find β is no larger than about 10⁻³ at 68% confidence. At that level, the T-duality model fits the data essentially as well as ΛCDM, with only a marginal statistical preference for the standard model. The paper establishes that any quantum-gravity-induced departure from ΛCDM in the late-time expansion is currently very small, and that future high-precision surveys will be needed to detect it.","feed_headline":"Zero-point length cosmology capped below 0.1% of ΛCDM","feed_subtitle":"Supernovae, BAO, and Hubble data set tight upper bounds on the T-duality parameter β, leaving standard cosmology intact.","key_machinery":"The key ingredient is the modified horizon entropy dS_h = 2πR(1 + l0²/R²)^(−3/2) dR, which follows from a T-duality-inspired zero-point length in the gravitational potential. Applying the first law of thermodynamics at the apparent horizon of a flat FRW universe produces the modified Friedmann equation H² − αH⁴ ≈ (8π/3)ρ + Λ/3, with α = 3l0²/4. The dimensionless parameter β = αH0² = (3/4)l0²H0² carries the deviation from ΛCDM, entering the Hubble function as H²/H0² = D(z)[1 + β D(z)] to first order, where D(z) = Ωm0(1+z)³ + Ωr0(1+z)⁴ + ΩΛ0. This β is the parameter the paper constrains with data.","core_discovery":"The central claim is that the string T-duality modified cosmology, with its zero-point length parameter β, is observationally viable but tightly bounded: a joint Bayesian analysis of late-time datasets gives β ≲ 10⁻³ (68% C.L.) when baryon acoustic oscillations are included, and the AIC model comparison shows the T-duality model and ΛCDM are statistically equivalent, with at most weak evidence favoring ΛCDM. This is the first quantitative observational constraint on this T-duality-inspired framework.","pith_inferences":["The paper leaves open whether the likelihood uses the exact Hubble solution (13) or the first-order expansion (14); for β near the reported upper bound, the square root in (13) becomes imaginary at the highest GRB redshifts, so the choice could bias the bound—this is an editorial concern, not a claim of the paper.","A natural extension is to apply the same observational framework to early-time data (CMB anisotropies, primordial gravitational waves), where the zero-point length is expected to leave stronger imprints than in late-time expansion.","The reported bound could be sharpened by replacing the GRB sample with a better-calibrated high-redshift distance indicator, since the current GRB data do not improve the constraint despite reaching z ≈ 8.","If future surveys push the β bound below ~10⁻⁴, the T-duality model would be effectively ruled out as a late-time modification, leaving only early-universe probes as viable tests."],"forward_implications":["If β is truly below ~10⁻³, the late-time expansion history of the universe is indistinguishable from ΛCDM with the current generation of distance and Hubble measurements.","The same zero-point length correction, which shows up as a subdominant late-time effect, could be far more visible in early-universe observables such as the CMB or primordial gravitational waves, providing a complementary test.","The bound on β translates directly into a bound on the zero-point length l0 ≈ (4β/3)^(1/2)/H0, connecting a cosmological measurement to the scale of spacetime discreteness.","The statistical equivalence of the two models means that adding more data of the same type will not by itself sharpen the constraint much; qualitatively new probes are needed to detect the correction.","The model can serve as a template for testing other minimal-length or quantum-gravity-inspired cosmologies against late-time data using the same Bayesian pipeline."],"fun_headline_variants":["T-duality cosmology passes tests, β below 10⁻³","String T-duality fits data, but ΛCDM still preferred","Zero-point length effect capped at 0.1% of standard model","First bounds on T-duality cosmology: β≤10⁻³","T-duality inspired cosmology passes cosmic tests"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The exact Hubble solution (13) is assumed to be valid over the entire redshift range of the data, including gamma-ray bursts at z ≈ 8.1, but for β near the reported upper bound the square root in that solution becomes imaginary at high redshift, so the analysis must either rely on the first-order expansion (14) or impose a hard existence cutoff—and the paper does not state which.","fun_headline_variants_meta":{"raw":{"variants":["T-duality cosmology passes tests, β below 10⁻³","String T-duality fits data, but ΛCDM still preferred","Zero-point length effect capped at 0.1% of standard model","First bounds on T-duality cosmology: β≤10⁻³","T-duality inspired cosmology passes cosmic tests"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000883,"raw_usage":{"total_tokens":3657,"prompt_tokens":753,"completion_tokens":2904,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":497,"completion_tokens_details":{"reasoning_tokens":2817}},"tokens_in":497,"tokens_out":2904,"duration_ms":17830,"temperature":1.0,"reasoning_tokens":2817,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-04T09:15:55.357833+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Recompute the distance modulus of the highest-redshift GRB (z ≈ 8.1) using the exact solution (13) versus the first-order expansion (14) at β = 0.003; if the two predictions differ by more than the reported measurement uncertainty, the upper bound on β depends on which expression was used in the likelihood, and the constraint is not purely data-driven.","supporting_citations":[],"review_version":1}