{"id":"6f34dd06-81bd-4e28-bf74-b2d0787e56e3","arxiv_id":"1908.08950","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"The authors show that cross-correlating GW source convergence with CMB lensing can measure a GR prediction (identical geodesics for GWs and photons) at high SNR for LISA-class detectors.","lead":"This paper proposes detecting the gravitational lensing of gravitational waves by cross-correlating gravitational-wave source distances with maps of the lensing of the cosmic microwave background. If the predicted correlation is seen by future detectors such as LISA or Einstein Telescope, it would test whether gravitational waves and photons follow identical paths through curved spacetime.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The estimator in Eqs. (17)–(18) assumes the distance-error field ε is uncorrelated with the CMB lensing field, but photometric redshifts and flux-limited EM-counterpart selection can make ε correlate with κ_CMB, biasing the claimed cross-correlation","rationale":"The reader's weakest-assumption is the same one I identify: the estimator in Eq. (18) is unbiased only if ε and κ_CMB are uncorrelated. The paper states this as an assumption rather than deriving it, and the assumed observing conditions (photometric redshifts for LISA counterparts, no EM counterparts for Cosmic Explorer) make correlated errors plausible through magnification bias and environment-dependent photo-z systematics. This is an external-validity concern rather than an internal inconsistency: Eqs. (14)–(15) give a standard Limber cross-spectrum, and the forecast machinery is coherent. But if correlated ε contributes to the recovered cross-spectrum, a nonzero measurement would not have the claimed interpretation as a clean GR verification. The proposed simulation test can settle the concern by injecting a density-dependent photo-z bias and checking whether the recovered C_l is biased above the noise. Given the reader's verdict is already CONDITIONAL, I do not recommend moving it; however, the overreach in the conclusion should be softened unless the test demonstrates negligible correlated-error bias.","tokens_in":10470,"tokens_out":7680,"duration_ms":93921,"concrete_test":"Build mock realizations (e.g., lognormal realizations of δ with the paper's W_gw and W_cmb kernels) with the fiducial LISA/CE source distributions. For each source, draw a true z, apply the assumed photometric error and a flux-limited EM selection with magnification μ≈1+2κ_gw, then add a parametrized density-dependent redshift bias δz/(1+z)=A κ_CMB(n). Reconstruct D̂_L via Eq. (17), estimate the cross-spectrum (18), and compare with the input C_l for A=0, 0.01, 0.1. If the recovered C_l shifts by more than the forecast statistical error in Fig. 3 for any A consistent with current photo-z calibration, the estimator is biased and the headline forecasts are optimistic.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The load-bearing step is the assertion after Eq. (18) that “the convergence field is uncorrelated with the error ϵ(n),” so ⟨ϵ κ_CMB⟩ drops out of E. In Eq. (17), ϵ = 1 − d_L/d_L^es contains the luminosity-distance measurement error, the photometric redshift error (σ_z/(1+z)=0.03 for LISA counterparts), and for Cosmic Explorer a 100% redshift uncertainty. These are not guaranteed to be independent of the matter distribution that lenses the CMB: flux-limited EM-counterpart selection is magnified by μ ≈ 1 + 2κ_gw, and photo-z residuals are known to correlate with galaxy environment, so E[ϵ | κ_CMB] need not vanish. Any such residual correlation contributes a term C_l^{ϵ κ_cmb} that is degenerate with the GR signal in Eq. (15). Additionally, the nonzero-correlation claim in the conclusion would not by itself “manifestly verify” identical geodesics unless the measured amplitude and redshift dependence are compared with the GR prediction, since many modified-gravity models also predict a nonzero GW–CMB correlation.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a new multi-messenger test of general relativity: reconstruct a weak-lensing convergence field from the luminosity-distance residuals of gravitational-wave (GW) sources relative to their electromagnetic (EM) redshifts, and cross-correlate this field with the CMB lensing convergence. The authors compute the theoretical cross-spectrum C_l^{kappa_gw kappa_cmb} under LCDM+GR using the Limber approximation and a nonlinear matter power spectrum (Eq. 15), estimate the covariance including GW shot noise and CMB lensing reconstruction noise (Eq. 19), and forecast cumulative SNRs for LISA and Cosmic Explorer as functions of source redshift, source number, and angular resolution (Fig. 3). The central claim is that a non-zero GW-CMB lensing correlation would verify a fundamental GR prediction that photons and gravitational waves follow the same geodesics, and that deviations from the predicted signal would probe modified gravity.","tokens_in":10702,"tokens_out":6288,"duration_ms":68846,"significance":"If the proposed measurement works as described, it would open a genuinely new observational window: no current test directly cross-correlates GW strain with CMB lensing to probe the equality of photon and graviton geodesics in the weakly perturbed cosmological metric. The theoretical framework is largely standard: Eq. (15) is the usual Limber-approximated cross-spectrum of two convergence fields, and Eq. (19) is a plausible diagonal covariance model. The paper is also commendably explicit about the approximations used (inspiral-only waveform, diagonal covariance, Limber approximation) and about the fact that Ngw and theta_min are treated as free parameters. The forecast is falsifiable in principle and would be a valuable target for LISA and next-generation ground-based detectors. The main issues are in the estimator: the treatment of the error term epsilon in Eq. (18), the reliance on an unpublished companion paper for the estimator, the unverified assumption of EM counterparts with photometric redshifts for LISA massive black hole binaries, and the concluding claim that a detected non-zero correlation would by itself 'manifestly verify' GR.","major_comments":[{"comment":"The estimator in Eq. (17) is ˆD_L = κ_gw + ε − εκ_gw, but the cross-correlation in Eq. (18) is written as ∫(ε + ˆκ_gw)ˆκ_CMB, silently dropping the product term −εκ_gw. Even if ⟨ε κ_CMB⟩ vanishes, the contribution −⟨εκ_gw κ_CMB⟩ does not automatically vanish, since ε and κ_gw probe the same line of sight; this missing term biases the estimated cross-correlation. More importantly, the assertion after Eq. (18) that ‘the convergence field is uncorrelated with the error ε(n)’ is an assumption rather than a demonstrated property. Flux-limited EM-counterpart selection is affected by lensing magnification, and photometric-redshift errors are known to correlate with galaxy environment, so ε can be correlated with the matter distribution that lenses the CMB. Such a residual correlation contributes a term C_l^{εκ_cmb} that is degenerate with the GR signal in Eq. (15). This is load-bearing because the whole claim of a clean cross-correlation measurement rests on this step.","section":"Estimator of the convergence field from GW strain, Eqs. (17)-(18)"},{"comment":"The conclusion states that ‘the existence of a non-zero correlation between these two signals will manifestly verify a fundamental prediction of general relativity.’ This overstates what the test can establish. First, the estimator in Eq. (17) uses the GR waveform relation of Eqs. (7)-(8) to convert measured strain into luminosity distance; any modified-gravity effect on wave propagation or on the distance–redshift relation would be absorbed into the reconstructed convergence field. Second, many modified-gravity theories predict a non-zero GW–CMB lensing correlation through altered lensing kernels, gravitational slip, or a modified growth rate, so a non-zero correlation is not by itself a GR discriminator. The paper should either weaken the claim to ‘consistent with the concordance GR prediction’ or specify how amplitude and redshift dependence would be compared with Eq. (15) to distinguish GR from alternatives.","section":"Conclusion"},{"comment":"The forecast treats Ngw and θ_min as free parameters and assumes that LISA massive black hole binaries have EM counterparts with σ_z/(1+z) = 0.03, while Cosmic Explorer sources have 100% redshift uncertainty. These assumptions are observationally unverified, and the projected SNR depends sensitively on them. The paper would be stronger if the source redshift distribution dngw/dz used in the forecasts were stated explicitly and if the SNR were presented as a function of the assumed counterpart fraction and redshift-error model. As it stands, the ‘large measurable window’ in Fig. 3 is a conditional forecast, not a prediction, and the reader cannot reproduce the cumulative SNR from the information given.","section":"Forecast for LISA and Cosmic Explorer, Fig. 3"},{"comment":"The core estimator of the convergence field and the detailed form of the luminosity-distance noise σ_dl are taken from Ref. [32], which is described as ‘Submitted to MNRAS (2019)’ and is not available to the reader of this manuscript. Given that the estimator is the central new element of the paper, the derivation of Eq. (17) and the mapping between σ_dl, σ_b, and the statistics of ε in Eq. (20) should either be provided in an appendix or the paper should cite a published, publicly available version of the companion work.","section":"Eqs. (17) and (20), estimator and noise model"}],"minor_comments":[{"comment":"The conclusion refers to ‘the predicted signal ... shown in Fig. 1’, but the signal as a function of source redshift is shown in Fig. 2.","section":"Conclusion"},{"comment":"The localization suppression factor is typeset as ‘el2θ2min/8 ln 2’; this should presumably be exp(l^2 θ_min^2 / (8 ln 2)) or an equivalent explicitly written exponential.","section":"Eq. (20)"},{"comment":"The notation ˆκ_CMB is introduced without definition; it should be stated that this is the reconstructed CMB lensing convergence map with the associated reconstruction noise N_l^{κκ}.","section":"Eq. (18)"},{"comment":"The sentence ‘As the convergence field is uncorrelated with the error ε(n), the first term on the right hand side goes to zero’ would be clearer if it explicitly distinguished the statistical ensemble over noise and source properties from the ensemble over cosmic density fields.","section":"General"}],"recommendation":"major_revision","confidential_remarks":"The proposal is timely and conceptually interesting, and the theoretical signal computation is standard and reproducible in principle. The main reservations are about the estimator: the uncorrelated-error assumption is not justified, the product term in Eq. (17) is dropped without discussion, and the central derivation is deferred to a companion manuscript. These issues are fixable within the scope of the paper, but they require a substantial revision of the estimator section and a more careful interpretation of what a detected correlation would mean. The paper fits the journal’s scope well, provided the forecast is reframed as conditional on currently unverified EM-counterpart and redshift assumptions."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear Colleague,\n\nShort take: this paper proposes cross-correlating the weak-lensing convergence field reconstructed from GW luminosity distances with CMB lensing, framed as a GW-CMB-CMB bispectrum. The core idea is good and the forecast is a useful step, but the paper overclaims what a nonzero correlation proves and hand-waves a bias term that could contaminate the signal.\n\nWhat's actually new: the GW-CMB-CMB three-point framing and the SNR forecasts for LISA and Cosmic Explorer with different source counts and sky localization. The lensing kernels and cross-spectrum (Eq. 15) are standard, and the covariance (Eq. 19) is the usual diagonal approximation. The paper is honest about the inspiral-only waveform and Limber approximation. That's fine for a proposal.\n\nThe soft spots are real but manageable. The biggest one is the claim after Eq. (18) that the distance-error field epsilon is uncorrelated with the CMB lensing field. That is not obviously true: photometric redshift errors are known to correlate with environment, and flux-limited EM-counterpart selection is magnified by lensing. If epsilon correlates with kappa_CMB, you get a term C_l^{epsilon kappa_cmb} degenerate with the signal. The paper needs to either justify this more carefully or propagate the bias in the forecast. The conclusion's \"manifestly verify GR\" is too strong—many modified gravity models also predict a nonzero correlation, so only amplitude and shape comparison can distinguish. Also, the estimator in Eq. (17) is deferred to an unpublished companion paper; for a measurement proposal, that's a missing piece. The forecast treats Ngw and theta_min as free, and assumes LISA sources have EM counterparts with sigma_z/(1+z)=0.03, which is not established. These are acknowledged in the text, but they make the SNR numbers illustrative rather than definitive.\n\nWho gets value: people planning LISA and CE science, and anyone working on GW lensing or multi-messenger gravity tests. It's worth a serious referee, but the referee should push for a cleaner treatment of the noise correlation and a toned-down conclusion.\n\nRecommendation: accept for peer review, with major revisions expected.","headline":"A solid proposal for a GW-CMB lensing cross-correlation test, but the claim that a nonzero correlation manifestly verifies GR is too strong and the distance-error correlation assumption needs work.","tokens_in":11262,"tokens_out":1649,"would_cite":true,"duration_ms":17025,"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":"Cross-correlating GW with CMB lensing can test whether gravity bends light and gravitational waves alike","keywords":["gravitational-wave lensing","CMB lensing","general relativity","modified gravity","multi-messenger astronomy","luminosity distance","cross-correlation","cosmology"],"falsifier":"If a future GW sample with the source counts, redshifts, and sky-localization errors assumed in the forecasts yields a cross-spectrum between the $\\hat{D}_L$ field and CMB lensing maps that is consistent with zero at cumulative SNR above 3, the paper's central prediction would be falsified; repeating the analysis with galaxy-lensing maps would provide a systematics check on the result.","tokens_in":10252,"feed_emoji":"🌌","tokens_out":11818,"duration_ms":100328,"temperature":0.7,"pith_summary":"This paper proposes a new way to test a fundamental prediction of general relativity: that photons and gravitational waves bend identically as they travel through the same perturbed spacetime. The idea is to estimate the weak-lensing convergence field from gravitational-wave luminosity distances and cross-correlate it with the lensing convergence map of the cosmic microwave background, producing a GW-CMB-CMB signal. Under GR and LCDM, the expected cross-correlation is nonzero and grows with source redshift, because high-redshift GW sources share more of their line of sight with CMB photons. The authors forecast that future detectors can measure this correlation with signal-to-noise above 3, which would manifestly verify the geodesic concordance of light and gravitational waves, while deviations would signal modified gravity.","feed_headline":"GW-CMB lensing cross-correlation could verify general relativity","feed_subtitle":"Future detectors can see this signal above SNR 3, testing whether photons and gravitational waves share geodesics.","key_machinery":"The load-bearing object is the GW luminosity-distance convergence estimator $\\hat{D}_L(\\hat{n}) = 1 - d_L(\\hat{n})/d_L^{\\rm est}(\\hat{n})$, which equals the lensing convergence $\\kappa_{\\rm gw}$ plus a distance-error term $\\epsilon$. Cross-correlating this estimator with a CMB lensing convergence map yields Eq. (18); because the error term is assumed uncorrelated with the CMB lensing field, the measured correlation isolates the common-lensing signal. The theoretical signal is the cross-spectrum $C_l^{\\kappa_{\\rm gw}\\kappa_{\\rm cmb}}$ of Eq. (15), built from the GW kernel $W_{\\rm gw}$ and CMB kernel $W_{\\rm cmb}$ and the nonlinear matter power spectrum $P_\\delta((l+1/2)/\\chi)$ in the Limber approximation.","core_discovery":"The central claim is that the lensing convergence field $\\kappa_{\\rm gw}$ imprinted on gravitational-wave luminosity distances by intervening matter can be estimated from the waveform alone, using the chirp-mass-independent luminosity distance estimator and an external redshift, and that this field has a predictable cross-power spectrum $C_l^{\\kappa_{\\rm gw}\\kappa_{\\rm cmb}}$ with CMB lensing. Because both photons and GWs propagate through the same gravitational potentials, GR predicts a nonzero correlation; the paper computes this signal in the Limber approximation from the product of the GW and CMB lensing kernels times the nonlinear matter power spectrum. The forecasts show LISA's massive black hole binaries with electromagnetic counterparts and Cosmic Explorer's stellar-mass binaries can detect the correlation with SNR above 3 for plausible source counts and sky localizations. A measured nonzero correlation verifies that GWs and photons follow identical geodesics, and a measured amplitude that deviates from the GR+LCDM prediction would probe modified gravity, graviton mass, extra dimensions, and differences between the scalar potentials $\\Phi$ and $\\Psi$.","pith_inferences":["If redshift errors or electromagnetic-counterpart selection effects correlate with the intervening matter distribution, the estimator would be biased; injecting such correlated errors into simulations would quantify how much bias can masquerade as a GR-violating signal.","The same convergence estimator could be cross-correlated with galaxy lensing or 21-cm lensing maps rather than CMB lensing, potentially providing higher-redshift overlap and independent systematics checks for stellar-mass GW sources without electromagnetic counterparts.","A tomographic version using GW source redshift bins could measure the growth of the lensing kernels directly, effectively mapping the matter distribution with gravitational waves as a probe, which the paper leaves implicit.","Because the GW-CMB cross-spectrum is sensitive to the same matter distribution as CMB lensing reconstruction, a discrepancy between the measured and predicted correlation could flag systematics in the CMB lensing maps themselves."],"forward_implications":["A nonzero GW-CMB lensing cross-correlation measured at SNR above 3 would directly verify general relativity's prediction that photons and gravitational waves travel on identical geodesics.","The measured amplitude of the cross-spectrum as a function of source redshift can constrain modified-gravity parameters and the relation between the scalar potentials $\\Phi$ and $\\Psi$.","The cross-correlation between GW strain and CMB lensing is effectively a three-point function of two CMB fields and the GW strain, going beyond the standard luminosity-distance-redshift test of background cosmology.","The same estimator can be applied to neutron star binaries, black hole-neutron star binaries, and to galaxy-survey lensing maps, widening the test to other GW sources and density tracers.","Removing the weak-lensing contribution from GW signals (delensing) reduces the uncertainty in GW source parameters and sharpens gravitational-wave tests of modified gravity."],"supporting_citations":[{"why":"It supplies the general-relativistic result that weak lensing affects photons and gravitational waves identically, which is the prediction the test targets.","marker":"[1]"},{"why":"It provides the quadratic minimum-variance estimator and reconstruction noise used to build the CMB lensing convergence map.","marker":"[16]"},{"why":"It gives the Newtonian inspiral waveform used to compute the GW signal-to-noise and luminosity-distance errors in the forecasts.","marker":"[19]"},{"why":"It establishes the chirp-mass-independent luminosity distance estimator that the convergence estimator relies on.","marker":"[22]"},{"why":"It derives the effect of cosmological perturbations on GW strain and introduces the GW convergence field $\\kappa_{\\rm gw}$.","marker":"[25]"},{"why":"It is the companion paper with the detailed derivation of the $\\hat{D}_L$ estimator of the convergence field from GW luminosity distances.","marker":"[32]"},{"why":"It gathers the arguments and simulations that LISA massive black hole binaries have electromagnetic counterparts, providing the redshifts needed for the estimator.","marker":"[33–38]"},{"why":"It supplies the LISA sensitivity curve used in the signal-to-noise forecasts.","marker":"[51]"},{"why":"It provides the merger-rate estimates used to set the number of Cosmic Explorer sources in the forecast.","marker":"[55]"}],"fun_headline_variants":["Weak lensing ties GW and CMB to test gravity","Test gravity with GW and CMB lensing cross-power","A gravitational test from cross-lensing of GW and CMB","GW-CMB lensing: a new probe of gravity's geodesics","Future detectors can test gravity via GW-CMB lensing"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The scheme assumes that errors in the estimated luminosity distance—from redshift uncertainties, distance measurement noise, and cosmological parameter errors—are completely uncorrelated with the CMB lensing convergence field, and that the GW sources used have reliable redshifts; if those errors correlate with the intervening matter distribution, the measured cross-correlation would be biased.","fun_headline_variants_meta":{"raw":{"variants":["Weak lensing ties GW and CMB to test gravity","Test gravity with GW and CMB lensing cross-power","A gravitational test from cross-lensing of GW and CMB","GW-CMB lensing: a new probe of gravity's geodesics","Future detectors can test gravity via GW-CMB lensing"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000715,"raw_usage":{"total_tokens":3169,"prompt_tokens":853,"completion_tokens":2316,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":469,"completion_tokens_details":{"reasoning_tokens":2231}},"tokens_in":469,"tokens_out":2316,"duration_ms":15426,"temperature":1.0,"reasoning_tokens":2231,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T11:25:49.509705+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"If a future GW sample with the source counts, redshifts, and sky-localization errors assumed in the forecasts yields a cross-spectrum between the $\\hat{D}_L$ field and CMB lensing maps that is consistent with zero at cumulative SNR above 3, the paper's central prediction would be falsified; repeating the analysis with galaxy-lensing maps would provide a systematics check on the result.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"It supplies the general-relativistic result that weak lensing affects photons and gravitational waves identically, which is the prediction the test targets."},{"cited_title":"Baker, E","cited_arxiv_id":null,"evidence_quote":"It provides the quadratic minimum-variance estimator and reconstruction noise used to build the CMB lensing convergence map."},{"cited_title":"Mukherjee, B","cited_arxiv_id":null,"evidence_quote":"It is the companion paper with the detailed derivation of the $\\hat{D}_L$ estimator of the convergence field from GW luminosity distances."}],"review_version":1}