{"id":"48955c80-7ab4-420c-8db9-4975ec03cf56","arxiv_id":"2412.15654","paper_version":2,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For a scale-invariant inflationary gravitational wave background with negligible primordial non-Gaussianity, the gravitational wave energy density perturbation is lognormally distributed, δGW = A e^{-ζ} - 1, yielding a specific connected three-point function and an effective f_NL^GW = f_NL - 5/6.","lead":"This paper derives a non-perturbative (all-orders) description of anisotropies in the cosmological gravitational wave background, showing that for a scale-invariant inflationary spectrum the energy density fluctuations are lognormally distributed. It matters because future gravitational wave observatories aim to measure these anisotropies, and nonlinear corrections could bias or extend the interpretation of the data.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Internal factor-of-4 inconsistency: with n_gwb=0, Eq. (2.10) implies δGW = A e^{-4ζ}-1, not A e^{-ζ}-1, so the exact bispectrum (3.12) and f_NL^GW = f_NL - 5/6 are incorrect as stated.","rationale":"The reader's verdict is CONDITIONAL because of the unproven exponential ansatz in Eq. (2.10). That is a legitimate concern, but the most load-bearing problem is more concrete and more severe: even accepting the ansatz, the paper's final expressions do not follow from its own definitions. The factor 4 arises from the energy weighting of the graviton distribution function: f ∝ q^{n-4} makes δGW = e^{(4-n)Γ}-1, and for n=0 this is e^{4Γ}-1. With Γ0=-ζ, the observable is A e^{-4ζ}-1, not A e^{-ζ}-1. The paper's path from Γ0 to δGW in Eq. (3.7) drops this density-weighting factor. This is not a convention issue: the linear coefficient of δGW is -4ζ, as seen directly from Eq. (2.19) with the SW factor from Eq. (2.30). The claimed exact three-point function (3.12) and the f_NL shift (3.15) are therefore incorrect as stated. The lognormal structure survives, but with a rescaled exponent, so the framework is salvageable by replacing -ζ with -4ζ (and correspondingly adjusting A and all correlation functions). However, the paper as written contains an internal inconsistency in its central result, so the verdict should move from CONDITIONAL to REJECT.","tokens_in":10796,"tokens_out":19691,"duration_ms":127687,"concrete_test":"Set Φ=Ψ=constant and n_gwb=0. Compute δGW at linear order from Eqs. (2.19) and (2.30): δGW = 2Ψ+4Φ. Using Eq. (3.2), ζ = -Φ-Ψ/2, so δGW_lin = -4ζ. Now expand Eq. (3.7) with ζ = -3Φ/2: δGW_lin = -ζ = 3Φ/2, which differs by a factor of 4. Equivalently, re-derive Eq. (3.11) directly from δGW = A e^{4Γ0}-1 with Γ0=-ζ; the two-point function becomes e^{16⟨ζ1ζ2⟩}-1, not e^{⟨ζ1ζ2⟩}-1. If the independent expansion reproduces -4ζ, Eq. (3.7) and all subsequent formulas must be revised.","verdict_should_be":"REJECT","load_bearing_attack":"The paper's own definitions force a factor 4 in the exponent of δGW that is missing from Eq. (3.7). From Eq. (2.10) with n_gwb=0, fGW ∝ q^{-4} e^{4Γ}, so the energy-density perturbation defined in Eq. (2.6) is δGW = f/f̄ - 1 = e^{4Γ}/⟨e^{4Γ}⟩ - 1. Combining Eqs. (2.21) and (2.31) gives Γ0 = -ζ, so the physical anisotropic energy density must be δGW = A e^{-4ζ} - 1 with A = 1/⟨e^{-4ζ}⟩. The paper instead writes δGW = A e^{-ζ} - 1 with A = 1/⟨e^{-ζ}⟩ (Eq. 3.7), dropping the factor 4. This is not a harmless convention: the linear limit of Eqs. (2.19)+(2.30) with Φ=Ψ yields δGW_lin = 2Ψ+4Φ = -4ζ, while expanding Eq. (3.7) gives δGW_lin = -ζ. Consequently the claimed exact two- and three-point functions (3.11)-(3.12) are those of e^{-ζ}, not of the defined δGW; under the assumed Γ0=-ζ they should read e^{16⟨ζ1ζ2⟩}-1 and the analogous 16-weighted combinations. The effective f_NL shift would be f_NL - 10/3, not f_NL - 5/6. This internal inconsistency undermines the central quantitative claims independently of whether the exponential ansatz is physically justified.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper presents a non-perturbative treatment of large-scale anisotropies of the cosmological gravitational wave background (CGWB). Assuming a power-law-exponential form for the graviton phase-space distribution function (Eq. 2.10), the authors derive the non-linear extension of the initial conditions and Sachs-Wolfe effect, express the result in terms of the comoving curvature perturbation ζ, and compute the two- and three-point correlation functions of the GW energy-density perturbation for a scale-invariant spectrum (n_gwb = 0) with negligible primordial non-Gaussianity. They obtain a lognormal distribution for δ_GW and an 'exact' bispectrum with effective f_NL^GW = f_NL - 5/6.","tokens_in":11179,"tokens_out":12546,"duration_ms":90583,"significance":"The geodesic and rescaling derivations in Section 2 are internally consistent, reduce to the known linear results of [32], and the step Γ0 = -ζ for n_gwb=0 is correct. If the final statistical results were correct, the paper would provide a useful non-perturbative benchmark for CGWB anisotropy searches. However, the central formula (Eq. 3.7) contains an internal factor-of-4 inconsistency that invalidates the reported correlation functions as they stand, and the exponential ansatz (Eq. 2.10) is assumed rather than derived, so the claimed 'exactness at any order' is conditional. The paper has no free parameters and its derivations are transparent, which is a strength.","major_comments":[{"comment":"The expression δ_GW = A e^{-ζ} - 1 contradicts the paper's own definitions. From Eq. (2.10) with n_gwb = 0, the distribution function is f_GW ∝ q^{-4} e^{4Γ}, so the fractional perturbation defined in Eq. (2.6) is δ_GW = e^{4Γ0}/⟨e^{4Γ0}⟩ - 1. Using Eq. (3.6) (Γ0 = -ζ) gives δ_GW = A e^{-4ζ} - 1 with A = 1/⟨e^{-4ζ}⟩, not e^{-ζ} as written. The linear limit of Eqs. (2.19) and (2.31) yields δ_GW_lin = -4ζ (for Φ=Ψ and no ISW), while Eq. (3.7) gives -ζ. Consequently, Eqs. (3.11)-(3.12) and (3.15) are the correlation functions of e^{-ζ}, not of the defined δ_GW; the correct expressions have e^{16⟨ζ_i ζ_j⟩} and an effective shift f_NL - 10/3.","section":"Section 3, Eq. (3.7)"},{"comment":"The central statistical results rest on the assumed exponential form of the graviton distribution function, which is introduced as an ansatz rather than derived from the theory. The paper later calls the three-point function 'exact and hold at any order' (Introduction), but this exactness is only relative to the ansatz. If the true nonlinear distribution deviates from Eq. (2.10), the lognormal property and the specific bispectrum do not follow. The authors should either justify the ansatz from the structure of the nonlinear theory or explicitly state that the results are conditional on this phenomenological model and temper the 'exact' claims.","section":"Section 2, Eq. (2.10)"}],"minor_comments":[{"comment":"The exponent is written as -Γ_in(n_gwb-4); since -Γ_in(n_gwb-4) = (4-n_gwb)Γ_in, using the latter form would make the expression visibly consistent with Eq. (2.10).","section":"Section 2, Eq. (2.21)"},{"comment":"After the factor correction, the normalization constant should be explicitly A = 1/⟨e^{-4ζ}⟩, and the text should note that the distribution is lognormal with variance 16⟨ζ^2⟩.","section":"Section 3, Eq. (3.7)"},{"comment":"The statement that the CGWB SW effect is 'completely insensitive to gravitational slip' is specific to n_gwb = 0; for general n_gwb the kernel K does not vanish (Eq. 3.5). The text should state this restriction.","section":"Conclusion"},{"comment":"Intermittency is invoked as a consequence of lognormality but no quantitative metric is given (e.g., a one-point PDF or a higher-moment ratio); adding one would strengthen the claim.","section":"Introduction"}],"recommendation":"major_revision","confidential_remarks":"The factor-of-4 error appears to be a single slip that propagates through Section 3; it is fixable but changes the paper's headline numbers. The deeper concern is the unproven exponential ansatz: the paper's 'non-perturbative exact' results are exact consequences of that ansatz, not of the underlying theory. I recommend major revision; the authors should correct the factor and clearly scope the claims."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Bottom line: there is real content here, but the central formula in Section 3 is off by a factor of 4, and the reader's report missed it. The paper genuinely derives a non-perturbative Sachs-Wolfe expression for CGWB anisotropies, shows the non-local slip kernel drops out when ngwb=0, and obtains a lognormal energy-density perturbation. That part is new and worth engaging. The problem is that eq. (3.7) writes delta_GW = A e^{-zeta} - 1, whereas the paper's own definitions (eqs. 2.6, 2.10, 2.21, 2.31) give delta_GW = A e^{-4 zeta} - 1. The factor (4-ngwb) is dropped when going from Gamma0 to delta_GW. This is not a convention: expanding the correct expression at ngwb=0 gives linear delta_GW = -4 zeta, while eq. (3.7) gives -zeta. Consequently the claimed exact two- and three-point functions (3.11)-(3.12) and f_NL^GW = f_NL - 5/6 are not the correlators of the delta_GW they defined; the correct versions carry powers of 16 in the exponents and shift f_NL by -10/3. The structure survives—delta_GW is still lognormal, the kernel still drops—but the quantitative claims as printed are wrong.\n\nSoft spot number two is the one the reader flagged: the exponential ansatz (eq. 2.10) is assumed, not derived. The paper is honest about this, but it means 'exact at any order' is conditional on a functional form motivated by the linear result. That was already a reason for a conditional verdict. The factor-4 error makes the current version unsuited for publication as is.\n\nWhat the paper does well: the geodesic and rescaling derivations are clear, reduce to published first-order results, and the application of the CMB path-integral machinery to the CGWB is a legitimate advance, especially the demonstration that the slip kernel K drops out for scale-invariant spectra beyond the squeezed limit.\n\nBottom line for you: this deserves a serious referee, not a desk reject, but it needs a revision that propagates the factor (4-ngwb) through Section 3. I would not cite the current version; I would cite a corrected one. For your reading group, it might be a good case study in how a simple factor slip can change the headline numbers.","headline":"A real all-orders CGWB result, but Section 3 drops a factor (4-ngwb) and the headline bispectrum numbers are wrong as printed.","tokens_in":11697,"tokens_out":11270,"would_cite":false,"duration_ms":92391,"reading_group":"yes","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 derives a non-perturbative Sachs-Wolfe mapping for the cosmological gravitational wave background and shows that, for a scale-invariant inflationary spectrum with negligible primordial non-Gaussianity, the energy density…","keywords":["cosmological gravitational wave background","gravitational wave anisotropies","non-perturbative Sachs-Wolfe effect","lognormal field","gravitational wave bispectrum","primordial non-Gaussianity","inflationary tensor modes","intermittency"],"falsifier":"Derive the graviton distribution function to all orders from the inflationary tensor-scalar system: if the momentum dependence is not exactly $(q e^{-\\Gamma})^{n_{\\rm gwb}-4}$, the predicted lognormal $\\delta_{\\rm GW}$ and exact three-point function fail. Observationally, measure the CGWB bispectrum on large angular scales with a future detector network; for $n_{\\rm gwb}=0$ it must have the shape and amplitude set by $W^{(3)}(x_1,x_2,x_3)$ with effective $f_{\\rm NL}^{\\rm GW} = f_{\\rm NL} - 5/6$, so a measurement with a clearly different amplitude or angular dependence would rule out the claim.","tokens_in":10600,"feed_emoji":"🌌","tokens_out":13731,"duration_ms":105543,"temperature":0.7,"pith_summary":"This paper derives a non-perturbative description of large-scale anisotropies in the cosmological gravitational wave background (CGWB), going beyond the first-order Boltzmann treatment. The central claim is that for a background generated by quantum tensor fluctuations during inflation, with a scale-invariant spectrum ($n_{\\rm gwb}=0$) and negligible primordial non-Gaussianity, the large-scale anisotropy is exactly $\\Gamma_0 = -\\zeta$, so the energy density perturbation $\\delta_{\\rm GW} = A e^{-\\zeta} - 1$ (with $A$ fixed by the zero-mean condition) is lognormally distributed whenever the curvature perturbation $\\zeta$ is Gaussian. This gives an exact connected three-point function $W^{(3)}(x_1,x_2,x_3) = e^{\\langle \\zeta_1 \\zeta_2 \\rangle + \\langle \\zeta_2 \\zeta_3 \\rangle + \\langle \\zeta_1 \\zeta_3 \\rangle} - e^{\\langle \\zeta_1 \\zeta_2 \\rangle} - e^{\\langle \\zeta_2 \\zeta_3 \\rangle} - e^{\\langle \\zeta_1 \\zeta_3 \\rangle} + 2$ and an effective nonlinearity parameter $f_{\\rm NL}^{\\rm GW} = f_{\\rm NL} - 5/6$. Because the derivation is non-perturbative, the result holds at every order in the scalar perturbations rather than only at leading order. A sympathetic reader cares because CGWB anisotropies are a future observable and their non-Gaussian statistics can discriminate between inflationary and astrophysical backgrounds.","feed_headline":"Gravitational-wave background anisotropies become lognormal, exactly","feed_subtitle":"For a scale-invariant inflationary background the exact bispectrum is fixed; future GW detectors can look for it.","key_machinery":"The machinery is the graviton phase-space distribution function written in the all-orders form $f_{\\rm GW}(q,\\Gamma) \\propto (q e^{-\\Gamma})^{n_{\\rm gwb}-4}$, where $\\Gamma$ is the dimensionless graviton energy perturbation and $n_{\\rm gwb}$ the spectral index of the background. Liouville's theorem (the collisionless Boltzmann equation, $df_{\\rm GW}/d\\eta=0$) makes the distribution constant along null geodesics, so the observed anisotropy is fixed by the ratio of emitted and observed comoving momenta. That ratio is computed non-perturbatively by absorbing superhorizon scalar perturbations into a redefined scale factor and time, $\\tilde a = e^{-\\Phi}a$, $d\\tilde\\eta = e^{\\Psi+\\Phi}d\\eta$, and integrating the time component of the graviton geodesic equation, giving $\\Gamma_0$ as an exponentiated combination of the potentials $\\Phi$, $\\Psi$. The correlation functions then follow from the path-integral generating functional $Z[J] = \\int \\mathcal{D}[\\zeta] P[\\zeta] e^{i\\int d\\vec x\\, J(\\vec x)(A e^{-\\zeta(\\vec x)}-1)}$, which turns the lognormal form of $\\delta_{\\rm GW}$ into closed-form expressions for all $n$-point functions via the standard identity for exponentials of a Gaussian field.","core_discovery":"The paper establishes that the linear relation between the observed graviton energy-density perturbation and the metric perturbations is only the first term of an exact all-orders mapping. Using the collisionless Boltzmann equation and the assumed form $f_{\\rm GW}(q,\\Gamma) \\propto (q e^{-\\Gamma})^{n_{\\rm gwb}-4}$ for the graviton distribution, it obtains the non-perturbative initial conditions and Sachs-Wolfe effect $\\Gamma_0 = \\frac{-2\\Psi(\\eta_{\\rm in})+4\\Phi(\\eta_{\\rm in})}{4-n_{\\rm gwb}(q)} + \\Psi(\\eta_{\\rm in}) + \\int_{\\eta_{\\rm in}}^{\\eta_0} d\\eta\\,(\\Phi' + \\Psi')$. Rewritten in terms of the comoving curvature perturbation, $\\Gamma_0 = -\\frac{2}{3}\\frac{6-n_{\\rm gwb}}{4-n_{\\rm gwb}}\\zeta - \\frac{2}{3}\\frac{n_{\\rm gwb}}{4-n_{\\rm gwb}}K$, with $K$ a non-local kernel coming from gravitational slip. For the scale-invariant inflationary case $n_{\\rm gwb}=0$ the kernel vanishes identically and $\\Gamma_0 = -\\zeta$ for every configuration, not only in the squeezed limit. Consequently $\\delta_{\\rm GW} = A e^{-\\zeta} - 1$ is lognormal if $\\zeta$ is Gaussian, the connected two-point function is $e^{\\langle\\zeta_1\\zeta_2\\rangle}-1$, the connected three-point function is $e^{\\langle\\zeta_1\\zeta_2\\rangle+\\langle\\zeta_2\\zeta_3\\rangle+\\langle\\zeta_1\\zeta_3\\rangle} - e^{\\langle\\zeta_1\\zeta_2\\rangle} - e^{\\langle\\zeta_2\\zeta_3\\rangle} - e^{\\langle\\zeta_1\\zeta_3\\rangle} + 2$, and the effective nonlinearity parameter is $f_{\\rm NL}^{\\rm GW} = f_{\\rm NL} - 5/6$; hence even a vanishing primordial $f_{\\rm NL}$ leaves a residual bispectrum from post-inflationary nonlinear evolution.","pith_inferences":["Not stated in the paper, but a direct extension: the ansatz for the graviton distribution is the hinge, so a first-principles derivation of $f_{\\rm GW}$ from inflationary tensor production would settle whether the lognormal prediction survives; a second-order mismatch would leave only the leading-order results intact.","A further extension: the same exponentiation applied to the integrated Sachs-Wolfe term would restore sensitivity to anisotropic stress and give an additional large-scale observable.","The CMB temperature and the CGWB energy density are both non-linear functions of the same $\\zeta$; although the paper does not say so, their large-scale bispectra should then share a fixed relation that a joint measurement could test."],"forward_implications":["For a scale-invariant spectrum ($n_{\\rm gwb}=0$) the non-local kernel $K$ drops out, so the large-scale Sachs-Wolfe anisotropy is exactly $\\Gamma_0=-\\zeta$ and is insensitive to gravitational slip for every configuration; slip effects are confined to the integrated Sachs-Wolfe piece.","The lognormal form fixes all connected correlators in terms of the curvature two-point function; the exact three-point function is $W^{(3)}(x_1,x_2,x_3) = e^{\\langle\\zeta_1\\zeta_2\\rangle+\\langle\\zeta_2\\zeta_3\\rangle+\\langle\\zeta_1\\zeta_3\\rangle} - e^{\\langle\\zeta_1\\zeta_2\\rangle} - e^{\\langle\\zeta_2\\zeta_3\\rangle} - e^{\\langle\\zeta_1\\zeta_3\\rangle} + 2$.","The effective bispectrum nonlinearity parameter is $f_{\\rm NL}^{\\rm GW} = f_{\\rm NL} - 5/6$, so the CGWB bispectrum has an irreducible contribution of $-5/6$ even when primordial non-Gaussianity is absent.","Lognormal statistics imply intermittency: the GW energy density develops rare high-density spots separated by large underdense regions, giving the cosmological signal a clumpy morphology distinct from the smoother astrophysical background.","The non-perturbative initial conditions and Sachs-Wolfe mapping are valid at all orders in scalar perturbations and provide the template for extending the same treatment to the integrated Sachs-Wolfe effect, lensing, and Shapiro time delays."],"supporting_citations":[{"why":"Supplies the identity used here for the n-point functions of exponentials of a Gaussian random field.","marker":"[27]"},{"why":"Provides the path-integral generating functional method that turns the lognormal form into correlation functions.","marker":"[43, 44]"},{"why":"Introduces the rescaling of large-scale scalar perturbations into the metric, the technique the paper generalizes from CMB to gravitons.","marker":"[25, 26]"},{"why":"Gives the linear inflationary initial conditions for the CGWB that this work extends to all orders.","marker":"[31, 32]"},{"why":"Demonstrates the Liouville/geodesic treatment of second-order CMB anisotropies on which the graviton calculation is modelled.","marker":"[23, 24]"},{"why":"Motivates the power-law frequency profile of the graviton distribution used in the ansatz.","marker":"[30]"},{"why":"Supplies the lognormal statistics framework used to identify the distribution of the GW energy density perturbation.","marker":"[45, 46]"},{"why":"Defines the effective nonlinearity parameter f_NL, which the paper generalizes to f_NL^GW.","marker":"[50, 51]"}],"fun_headline_variants":["Exact non-linear CGWB anisotropies turn lognormal for scale-invariant case","Gravitational wave background anisotropies become exactly lognormal","Non-linear effects make CGWB anisotropies lognormal exactly","Lognormal GW background from exact nonlinear Sachs-Wolfe","Exact CGWB bispectrum shows residual signal without primordial NG"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The paper's \"exact at any order\" conclusions stand on the assumption, stated as equation (2.10), that the full graviton distribution function has the power-law-exponential form $f_{\\rm GW}(q,\\Gamma) \\propto (q e^{-\\Gamma})^{n_{\\rm gwb}-4}$; this form is chosen to reproduce the linear result but is not derived from the underlying theory, and if the true distribution differs the lognormal energy density and the exact three-point function do not follow.","fun_headline_variants_meta":{"raw":{"variants":["Exact non-linear CGWB anisotropies turn lognormal for scale-invariant case","Gravitational wave background anisotropies become exactly lognormal","Non-linear effects make CGWB anisotropies lognormal exactly","Lognormal GW background from exact nonlinear Sachs-Wolfe","Exact CGWB bispectrum shows residual signal without primordial NG"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00077,"raw_usage":{"total_tokens":3518,"prompt_tokens":1162,"completion_tokens":2356,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":778,"completion_tokens_details":{"reasoning_tokens":2278}},"tokens_in":778,"tokens_out":2356,"duration_ms":14763,"temperature":1.0,"reasoning_tokens":2278,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T11:13:50.601954+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Derive the graviton distribution function to all orders from the inflationary tensor-scalar system: if the momentum dependence is not exactly $(q e^{-\\Gamma})^{n_{\\rm gwb}-4}$, the predicted lognormal $\\delta_{\\rm GW}$ and exact three-point function fail. Observationally, measure the CGWB bispectrum on large angular scales with a future detector network; for $n_{\\rm gwb}=0$ it must have the shape and amplitude set by $W^{(3)}(x_1,x_2,x_3)$ with effective $f_{\\rm NL}^{\\rm GW} = f_{\\rm NL} - 5/6$, so a measurement with a clearly different amplitude or angular dependence would rule out the claim.","supporting_citations":[{"cited_title":"Non-Gaussianity of Large-Scale Cosmic Microwave Background Anisotropies beyond Perturbation Theory,","cited_arxiv_id":null,"evidence_quote":"Supplies the identity used here for the n-point functions of exponentials of a Gaussian random field."},{"cited_title":"Gravitational waves from inflation in LISA: reconstruction pipeline and physics interpretation,","cited_arxiv_id":null,"evidence_quote":"Motivates the power-law frequency profile of the graviton distribution used in the ansatz."}],"review_version":1}