{"id":"78086b19-4749-4614-8884-3ea77b977bc6","arxiv_id":"1908.01361","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The transverse motion of a gravitational lens shifts the frequency of microlensed starlight, and measuring that shift could break the mass-distance-velocity degeneracy of microlensing events.","lead":"This paper proposes measuring a tiny shift in the color of starlight during gravitational microlensing to measure how fast the lensing object moves across the sky. If the shift can be seen, it would help measure the masses and velocities of dark, invisible objects like black holes and free-floating planets.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Eq. (6) drops a first-order observer Doppler term α(v_E−v_S)/c, so Eq. (17) does not isolate vLS even in the ideal S/N limit.","rationale":"I read the paper as a proposal to measure the lens-source transverse velocity through a frequency shift that is derived in the source rest frame and then compared with Earth-based spectrograph sensitivity. The algebraic path from the Lorentz transformations to Eqs. (6) and (17) is coherent within the stated frame, and the flux-weighted combination in Eq. (16) also follows. The reader's observability objection is real: the ESPRESSO 0.1 m/s limit corresponds to δν/ν ≈ 3×10^-10, about thirty times larger than the fiducial 1×10^-11 shift, and finite-source and blending effects are not quantified. I agree with that assessment. However, I found a more internal weakness: the paper's assertion that an observer moving relative to the source adds only O((v_E−v_S)^2) corrections is order-of-magnitude incorrect. The photon direction changes by the deflection angle α, so the standard Doppler term u·k contains α times the transverse observer velocity. This is first order in both α and u and is comparable to the claimed signal under the paper's own fiducial parameters. The concrete test of transforming Eq. (5) to the Earth frame would settle this analytically. If the extra term is present, Eq. (17) must be revised to include the known observer motion; otherwise fits to real data would confuse vLS with a component of v_E−v_S. The method is salvageable because the observer velocity is known to sufficient accuracy from Earth ephemerides and GAIA, so I do not recommend rejection. The correct response remains a conditional acceptance, with the added condition that the observer Doppler term be included and the detectability analysis be redone.","tokens_in":6189,"tokens_out":18571,"duration_ms":226506,"concrete_test":"Re-derive the observed frequency by Lorentz-transforming the photon four-momentum of Eq. (5) into an Earth frame moving with velocity u = v_E − v_S relative to the source frame, rather than assuming the observer is at rest in the source frame. Keeping all first-order terms in α and u, compare the resulting lensing-induced δν/ν with Eq. (6): if an additional term α(γ u_x cosΩ + u_z sinΩ) appears, then Eq. (6) is incomplete. Recompute Eq. (17) including this term and refit the fiducial numbers to check whether the recovered vLS shifts by an O(u) amount.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The derivation of Eq. (6) is performed in the source rest frame, and the paper then dismisses the observer's motion relative to that frame as an O((v_E−v_S)^2) correction. That dismissal is wrong. From Eq. (5), the photon direction in the source frame after scattering is k ≈ y + (−γα cosΩ)x + (−α sinΩ)z. For an observer moving with u = v_E − v_S relative to this frame, the special-relativistic Doppler factor is 1 − u·k/c to first order in u. The u·k term contains u_x(−γα cosΩ) + u_z(−α sinΩ), i.e. a first-order product of α and u, not a second-order term. With u_x ≈ 200 km/s for a bulge source, this term has the same magnitude as −γ vLS α cosΩ for the fiducial vLS = 200 km/s. Since α and Ω vary on the Einstein timescale, this extra term contributes to the time-dependent flux-weighted frequency shift in Eq. (17). Omitting it means the measured shift is a blend of vLS and the transverse observer velocity, and a fit to Eq. (17) alone would bias the inferred lens velocity by up to the relevant component of u. The correction is known in principle from GAIA proper motion and Earth ephemeris, so the method can be repaired, but the paper's central 'direct measurement of vLS' claim is not supported by the equations as written.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes that the transverse motion of a gravitational lens relative to the source produces a frequency shift in the light rays passing the lens, and that this shift, time-resolved during a microlensing event, can be used to measure the lens-source transverse velocity. The derivation starts from Lorentz transformations between the source rest frame and the lens frame, combines the standard Schwarzschild deflection angle with the lens motion, and arrives at Eq. (6) for the frequency shift and Eq. (17) for the flux-weighted, time-dependent shift of the two images. The paper then argues that combining this spectroscopic measurement with GAIA proper motions and space-based parallax would break the microlensing degeneracy and yield the lens mass and velocity for objects ranging from black holes to free-floating planets.","tokens_in":6435,"tokens_out":8048,"duration_ms":78008,"significance":"The Lorentz-transformation derivation in Eqs. (2)-(6) is self-contained, internally consistent, and free of fitted parameters. If the effect were observable, the proposed combination of a spectroscopic radial-velocity-like observable with GAIA astrometry and space-based parallax would be an interesting new route to breaking the microlensing degeneracy. However, the manuscript as written does not establish observability: the fiducial signal is roughly two orders of magnitude below the ESPRESSO sensitivity that the paper itself quotes, and a first-order observer Doppler term is omitted from the central formula. The physical idea is worth pursuing, but the quantitative claims in the abstract and conclusion are not supported by the current analysis.","major_comments":[{"comment":"The statement that observer motion relative to the source frame produces only an O((v_E-v_S)^2) correction is incorrect. In the source rest frame, after scattering the photon direction is k ≈ ŷ - γα cosΩ x̂ - α sinΩ ẑ. For an observer moving with velocity u = (v_E - v_S)/c relative to this frame, the measured frequency is E_obs = E''(1 - u·k) to first order in u. The terms u_x γα cosΩ and u_z α sinΩ are first order in both u and α, hence the same order as -vLS α cosΩ in Eq. (6) for u_x ~ 200 km/s. Since α and Ω vary on the Einstein timescale, these terms contribute to the time-dependent flux-weighted shift in Eq. (17). Thus Eq. (17) does not isolate vLS; fitting it alone would bias the inferred lens velocity by an amount comparable to the relevant component of u. Because u is known in principle from the Earth ephemeris and GAIA proper motions, the model can be repaired, but the derivation as written is not correct.","section":"Eqs. (4)-(6) and discussion after Eq. (6)"},{"comment":"The paper's own numbers do not support the claimed measurability. For M = 0.5 M_sun, D_s = 8.5 kpc, x = 0.1, and vLS = 200 km/s, the peak value of the time-dependent factor in Eq. (17) is at most about 0.35, giving a shift of roughly 2.6 × 10^-12, not the ~10^-11 stated in the text. This is about 100 times smaller than the ESPRESSO 0.1 m/s sensitivity (δν/ν ~ 3 × 10^-10) quoted by the authors. Reaching 3 × 10^-10 at the same geometry would require a lens mass of order 7000 M_sun, so the proposal does not cover the claimed range from black holes to free-floating planets. The analysis also omits photon noise, spectral-line broadening, source finite-size effects, and blending, each of which would further reduce the observable signal.","section":"Eq. (17), Fig. 3, and comparison with ESPRESSO"}],"minor_comments":[{"comment":"The formula is written in units with c = 1; in SI units it should read δν/ν = -4GM vLS cosΩ / (b c^3). Stating this explicitly would avoid dimensional confusion.","section":"Eq. (6)"},{"comment":"The abstract quotes a sensitivity of 10^-11, while the fiducial peak shift from Eq. (17) is a few times 10^-12; the numbers should be made consistent.","section":"Abstract and Fig. 3"},{"comment":"There are multiple typos: 'Paczysnki' for Paczynski, 'angel' for angle, 'loos energy' for loses energy, and a stray vertical bar in the sign discussion after Eq. (6). The manuscript should be carefully copyedited.","section":"General presentation"},{"comment":"The y-axis label should specify that the plotted quantity is the dimensionless δν/ν, and the figure caption should list the values of x, Ω0, and the lens mass used.","section":"Figure 3"}],"recommendation":"major_revision","confidential_remarks":"The core physical effect is not new (see Refs. [10], [12]-[16]), but the application to microlensing spectroscopy is a reasonable idea. The two major issues are fixable: the observer Doppler term is known and can be added, and the detectability claim can be re-scoped to high-mass lenses with a realistic signal-to-noise estimate. I would encourage the authors to address these points before publication; the present version overstates both the novelty and the observational reach."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Couple of things you should know about 1908.01361. The underlying physics is not new—Pyne & Birkinshaw and Kopeikin & Schäfer already have the frequency shift from a moving lens—but the application to microlensing, specifically the flux-weighted time-dependent formula in Eq (17), is a genuinely new idea. If it worked, it would give a direct velocity handle on the lens and help break the mass-distance-velocity degeneracy. The Lorentz-transformation derivation is clean and self-contained, and the paper cites the prior work honestly. Credit where due: the paper is a solid example of taking a known relativistic effect and turning it into a proposed observable.\n\nThat said, there is a real problem in the derivation. The paper performs the calculation in the source rest frame, then says the observer's motion relative to that frame contributes only at O((v_E−v_S)^2). That is not right. After scattering, the photon direction has transverse components of order α (the deflection angle), so an observer moving with velocity u sees a Doppler factor 1 − u·n/c, and u·n contains α u/c, which is first order in α and in u—same order as the α v_LS/c term the paper wants to measure. For bulge sources with u_x ~ 200 km/s, this term is comparable to the fiducial v_LS term. So Eq (17) does not isolate v_LS as written. The correction is correctable in principle using GAIA proper motions and Earth ephemeris, but the paper doesn't do it.\n\nThe second soft spot is detectability. The paper's own numbers put the fiducial shift at about 10^-11, while ESPRESSO's 0.1 m/s resolution corresponds to about 3×10^-10. That is a factor of 30 short. The paper says 'for massive lenses we expect detection' but does not quantify what mass is needed or how the finite-source and blending effects would affect the line profile. The point-source, Dirac-delta treatment is a toy model.\n\nAll in all: the central idea is worth a serious look, but the central claim is not supported by the equations as written. This paper deserves a referee—it is not a desk reject—but the referee should ask for a corrected treatment of the observer term and a realistic sensitivity estimate. I would not cite Eq (17) in its current form, though I would cite the paper as a proposed observable if it were fixed.","headline":"A promising new microlensing observable with a derivation flaw: the observer's transverse motion enters at the same order as the lens velocity, and the predicted signal is below current spectrograph sensitivity.","tokens_in":6995,"tokens_out":3460,"would_cite":false,"duration_ms":33726,"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":"A gravitational lens's sideways motion shifts the frequencies of its two images in opposite directions; the net flux-weighted shift is measurable and encodes the lens–source velocity.","keywords":["gravitational microlensing","frequency shift","transverse velocity","spectroscopic follow-up","Einstein crossing time","microlensing degeneracy","lens mass measurement","relativistic Doppler effect"],"falsifier":"Take a well-measured microlensing event with an independently known massive lens and observe its spectrum repeatedly across the Einstein crossing with a spectrograph stable to $\\delta\\nu/\\nu\\sim 10^{-11}$; the prediction is a specific time-dependent shift with sign and amplitude given by Eq. (17). If no such shift appears in a favorable event, or if its measured amplitude disagrees with the independently known lens-source velocity, the proposed observable is refuted.","tokens_in":5949,"feed_emoji":"🔭","tokens_out":18091,"duration_ms":162541,"temperature":0.7,"pith_summary":"The paper proposes a new observable in gravitational microlensing: a spectral frequency shift produced by the transverse motion of the lens. In the standard point-source treatment, the two images are shifted by equal and opposite amounts, but the flux-weighted blend moves by roughly $10^{-11}$ in $\\delta\\nu/\\nu$ for a typical $0.5\\,M_\\odot$ lens. The author derives this shift from relativistic frame changes around the deflector and shows that it carries the relative transverse velocity $v_{LS}$ of lens and source. If measurable, that velocity would break the long-standing degeneracy among mass, distance, and velocity in microlensing fits. Combined with astrometry and space-based parallax, the same observation would yield the mass and transverse velocity of lenses from stellar black holes to free-floating planets.","feed_headline":"Moving lenses shift starlight by one part in ten trillion","feed_subtitle":"Measuring that shift with astrometry and parallax yields lens mass and speed, from black holes to free-floating planets.","key_machinery":"The load-bearing object is the flux-weighted spectral barycenter shift of the two microlensed images. The derivation combines a Lorentz boost from the source frame into the moving lens frame (giving the photon an $x$-momentum $-\\gamma v_{LS}E$ before scattering), the standard Schwarzschild deflection kick $\\Delta p'=-\\alpha p_y(\\cos\\Omega,\\sin\\Omega)$, and a boost back to the source frame, which leaves an energy change proportional to $v_{LS}$. The two images have opposite shifts, so they cancel in the total flux; the observable survives only after weighting each image's shift by its magnification, as in Eq. (15). With the point-source image positions and magnifications, this reduces to the explicit time-dependent formula (17) in terms of the impact parameter $\\beta_0$, the trajectory angle $\\Omega_0$, and the Einstein time $t_E$.","core_discovery":"The central claim is that gravitational microlensing should produce a small, time-dependent frequency shift in the source's spectral lines, set by the lens's transverse velocity. Treating the light ray as a photon that is Lorentz-transformed into the moving lens frame, given a Schwarzschild deflection, and transformed back yields $\\delta\\nu/\\nu=-4GM v_{LS}\\cos\\Omega/b$ to first order in $v_{LS}$. Averaging this shift over the two images with their magnification weights gives Eq. (17), a closed-form expression that peaks at roughly $10^{-11}$ for a $0.5\\,M_\\odot$ lens at typical bulge distances. Since the Einstein-crossing-time parameter $t_E$ alone cannot separate mass, distance, and velocities, this spectroscopic shift adds a direct handle on $v_{LS}$, which is exactly the missing kinematic quantity. The paper therefore presents spectroscopy as a route to fully determining the lens parameters in microlensing events.","pith_inferences":["I infer that the same flux-weighted shift should also be present in high-magnification caustic-crossing events, where the relative weights of the two images change rapidly, giving a stronger and more distinctive time signature per unit signal.","I infer that even below the single-event detection threshold, stacking many events with similar geometry could recover a statistical average of $v_{LS}$, turning the effect into a population probe of dark lenses.","I infer that the physical mechanism is the gravitational analogue of a slingshot: photons passing ahead of the moving lens lose energy and those passing behind gain energy, so the sign of the shift around the event encodes which side of the lens the source is passing."],"forward_implications":["Spectroscopic follow-up of a microlensing event would supply the relative transverse velocity $v_{LS}$ directly, breaking the mass-distance-velocity degeneracy that single light curves cannot resolve.","Combined with the source proper motion from astrometry and with space-based parallax measurements, the new velocity would allow the lens mass and full transverse motion to be solved, applicable to black holes, neutron stars, brown dwarfs, and free-floating planets.","The predicted shift changes sign and amplitude during the event, so time-series spectroscopy offers an internal consistency check of the lensing geometry.","Because the amplitude scales as the square root of the lens mass, the most massive lenses are the best targets, bringing the effect within range of current high-resolution spectrographs."],"supporting_citations":[{"why":"Establishes gravitational microlensing as the observational setting and defines the Einstein crossing time that this paper tries to supplement.","marker":"[1]"},{"why":"Supplies the full general-relativistic treatment of photon energy change used for the frequency-shift derivation.","marker":"[10]"},{"why":"Provides the moving-lens correction to the deflection angle needed in the Lorentz-transformation calculation.","marker":"[12]"},{"why":"Gives the Shapiro-delay route to the same frequency-shift formula, confirming the effect from a different formalism.","marker":"[15]"},{"why":"Is the earlier extensive study of gravitational lensing by moving lenses that this work adapts to microlensing observables.","marker":"[16]"},{"why":"Reports the measured transverse-motion frequency shift in the Cassini radio link, the observational precedent for the effect.","marker":"[19]"},{"why":"Defines the ESPRESSO radial-velocity precision that determines whether the predicted shift is currently detectable.","marker":"[20]"},{"why":"Provides the space-based parallax measurements that, with the new shift, break the microlensing degeneracy.","marker":"[22]"},{"why":"Supplies the astrometric microlensing measurements from GAIA that give the source-proper-motion piece of the velocity solution.","marker":"[9]"},{"why":"Sets out the degeneracy problem in microlensing that the proposed spectroscopic measurement is designed to resolve.","marker":"[4]"}],"fun_headline_variants":["Microlensing frequency shift reveals lens speed","New way to weigh black holes and free-floating planets","Spectroscopy unlocks lens mass and velocity in microlensing","Measuring the kick: microlensing's frequency shift"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The weak point is observability: the whole proposal depends on the frequency shift surviving in real observations, but the predicted size is about $10^{-11}$, roughly thirty times smaller than the $0.1\\,\\mathrm{m/s}$ precision of today's best spectrographs, and the calculation assumes a point-like source with a single sharp spectral line.","fun_headline_variants_meta":{"raw":{"variants":["Microlensing frequency shift reveals lens speed","New way to weigh black holes and free-floating planets","Spectroscopy unlocks lens mass and velocity in microlensing","Measuring the kick: microlensing's frequency shift"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000518,"raw_usage":{"total_tokens":2502,"prompt_tokens":930,"completion_tokens":1572,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":546,"completion_tokens_details":{"reasoning_tokens":1509}},"tokens_in":546,"tokens_out":1572,"duration_ms":12118,"temperature":1.0,"reasoning_tokens":1509,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T15:15:37.403202+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a well-measured microlensing event with an independently known massive lens and observe its spectrum repeatedly across the Einstein crossing with a spectrograph stable to $\\delta\\nu/\\nu\\sim 10^{-11}$; the prediction is a specific time-dependent shift with sign and amplitude given by Eq. (17). If no such shift appears in a favorable event, or if its measured amplitude disagrees with the independently known lens-source velocity, the proposed observable is refuted.","supporting_citations":[{"cited_title":"the lens (i.e π/2 <|Ω| < π|) gains energy","cited_arxiv_id":null,"evidence_quote":"Establishes gravitational microlensing as the observational setting and defines the Einstein crossing time that this paper tries to supplement."},{"cited_title":"Dong et al","cited_arxiv_id":null,"evidence_quote":"Supplies the full general-relativistic treatment of photon energy change used for the frequency-shift derivation."},{"cited_title":"Jeong, C","cited_arxiv_id":null,"evidence_quote":"Provides the moving-lens correction to the deflection angle needed in the Lorentz-transformation calculation."},{"cited_title":"Heyrovsky, The Astrophysical Journal, 624, 28-33 (2005)","cited_arxiv_id":null,"evidence_quote":"Gives the Shapiro-delay route to the same frequency-shift formula, confirming the effect from a different formalism."},{"cited_title":"Baghram, N","cited_arxiv_id":null,"evidence_quote":"Is the earlier extensive study of gravitational lensing by moving lenses that this work adapts to microlensing observables."},{"cited_title":"Measuring Limb Darkening of Stars in high magnification Microlensing Events by the Finite Element Method","cited_arxiv_id":"1906.10589","evidence_quote":"Reports the measured transverse-motion frequency shift in the Cassini radio link, the observational precedent for the effect."},{"cited_title":"Bertotti, L","cited_arxiv_id":null,"evidence_quote":"Defines the ESPRESSO radial-velocity precision that determines whether the predicted shift is currently detectable."},{"cited_title":"Poleski, W","cited_arxiv_id":null,"evidence_quote":"Supplies the astrometric microlensing measurements from GAIA that give the source-proper-motion piece of the velocity solution."},{"cited_title":"Mollerach, E","cited_arxiv_id":null,"evidence_quote":"Sets out the degeneracy problem in microlensing that the proposed spectroscopic measurement is designed to resolve."}],"review_version":1}