{"id":"7e13a815-014a-4f03-8254-e7f75c39f4a3","arxiv_id":"2501.13030","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"A classical local gravitational interaction necessarily causes momentum diffusion in quantum matter, and the required minimum diffusion could be detected by a millikelvin torsion pendulum.","lead":"This paper argues that if gravity is classical rather than quantum, it must randomly jiggle the motion of any nearby object, and that this jiggle is in principle measurable with an extremely quiet torsion pendulum. It offers a way to test the nature of gravity without building macroscopic quantum superpositions.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The first-order PPT expansion cannot yield Eq. (19): the stated null vector is not null, and the Newtonian term vanishes on the true null space, so Eq. (23) currently lacks a valid derivation.","rationale":"The reader's weakest assumption concerned frequency independence of the diffusion coefficients γij. That is a real concern, but Appendix E explicitly relaxes it for a symmetric setup, so it is not the most load-bearing issue. The more serious problem is internal to the derivation of the central quantitative claim. The first-order Taylor expansion in Eq. (18) is the only step linking the physical requirement of no entanglement to the concrete bound Eq. (23). That step is faulty in two independent ways: the vector claimed to annihilate the zeroth-order PPT matrix does not do so, and the true null vectors see no first-order effect of the Newtonian interaction. Consequently, the K-dependent lower bound cannot be obtained from the stated calculation. This does not prove the bound false, and the paper may be salvageable with a higher-order or exact analysis of the PPT condition, but as it stands the experimental protocol is calibrated against an unproven inequality. The verdict should therefore be moved below conditional acceptance until the bound is re-derived and verified. The concrete numerical check of the explicit master equation for the symmetric diagonal case would settle whether the final inequality survives; if it survives, a revised derivation should be supplied and the paper can be reconsidered.","tokens_in":26323,"tokens_out":48293,"duration_ms":545737,"concrete_test":"Recompute the claimed null vector: evaluate z0†(I4+iΛJΛ)z0 for z0=(a,-b,ia,ib); it equals 4|a|^2, not zero. Then re-derive the no-entanglement condition to second order in ε by integrating Eq. (17) exactly for the symmetric diagonal case (γ11=γ22, γ33=γ44) with γ11+m^2Ω^2γ33 = Gm^2/(ℏd^3) and checking whether ΛV(ε)Λ+iJ/2 remains positive semidefinite for ε up to several decoherence times. If the PPT condition fails, Eq. (23) is false; if it holds, the first-order proof in Section IV must be replaced by a correct second-order argument before the experimental protocol can be regarded as justified.","verdict_should_be":"UNVERDICTED","load_bearing_attack":"The quantitative bound Eq. (23) is derived from a first-order Taylor expansion of the PPT condition (Eq. (18)). This expansion is not valid as written. With V0=I4/2 and the ordering c=(x1,x2,p1,p2), the zeroth-order matrix M0=I4+iΛJΛ has null vectors with p=-i x. For the vector z0=(a,-b,ia,ib) stated in the text, one finds z0†M0z0=4|a|^2, or 2|a|^2 with the 1/2 prefactor in Eq. (18), so it does not vanish identically; the claimed null vector has the wrong sign for the first momentum component. More importantly, on the true null space p=-i x the first-order contribution of the Newtonian interaction K x1 x2 vanishes: z†(JH_int-H_intJ)z/2 = -Re[x†A p] = 0 because x†Ax is real. Physically, an x1x2 coupling cannot change the covariance matrix at first order starting from the product ground state, since all initial cross-correlations are zero; entanglement generation is second order in time. Thus a first-order expansion cannot produce the K-dependent term on the left-hand side of Eq. (19). The bound in Eq. (23) may be true, but the published derivation does not establish it, and this bound is the quantitative basis for the proposed experimental test.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript argues that if gravity is classical in the LOCC sense, it must act as a stochastic, diffusive channel on quantum matter. For two harmonically trapped equal masses it writes a general Lindblad master equation, imposes that Newtonian evolution is recovered on classical states, and uses preservation of separability under partial transposition to derive a lower bound on momentum diffusion, Eq. (23). It then proposes a torsion-pendulum protocol at millikelvin temperatures, with parameters in Table I, to detect the predicted heating rate. The conceptual argument is that classical gravity must collapse spatial superpositions to avoid superluminal signaling, and that this collapse is necessarily diffusive by the theorem of ref. [84].","tokens_in":26633,"tokens_out":11198,"duration_ms":121058,"significance":"If valid, the paper would provide a genuinely new experimental route: the test target becomes diffusion of a classical macroscopic probe rather than gravitationally induced entanglement of prepared quantum states. The paper is clear about the logical direction (the bound is necessary for separability, not sufficient for classicality), and it presents a detailed, falsifiable experimental protocol with quantitative requirements. It also includes useful appendix material on linearization, noise conditions, and symmetries. However, the quantitative foundation, especially Eq. (23), is not established by the present derivation.","major_comments":[{"comment":"The derivation of Eq. (19) from the first-order expansion of the PPT condition is not valid as written. With the ordering c=(x1,x2,p1,p2) and M0=I4+iΛJΛ, the kernel of M0 is spanned by vectors of the form (a,-b,ia,-ib), not by the vector z0=(a,-b,ia,ib) stated in the text: for the printed vector one obtains z0†M0z0=2|b|^2, so the first term in Eq. (18) does not vanish. More importantly, even with the corrected null vector, the first-order contribution of the Newtonian term to z†(dV/dt)z vanishes: writing H_int=λ x1x2 with λ=K/(m√Ω1Ω2), the initial covariance derivative contains a block [[0,-A],[-A,0]] with A=[[0,1],[1,0]], and on the null space p=-ix one obtains an expression proportional to Re[x†Ax]=0. Thus a first-order expansion cannot produce the K-dependent term in Eq. (19); entanglement generation is quadratic in time. The bound in Eq. (23) therefore currently lacks a valid derivation and must be re-derived by a second-order expansion or by an independent argument.","section":"IV, Eq. (18)-(19)"},{"comment":"The transition from the setup-dependent inequality Eq. (22) to the universal bound Eq. (23) relies on the assumption that the gravitational diffusion coefficients γij are independent of the trapping frequencies and that, for equal masses, γ11=γ22 and γ33=γ44. This is introduced as \"a reasonable expectation\" but is not derived from the stated assumptions. If γij depend on Ω1 and Ω2, Eq. (22) only constrains the specific trap configuration and Eq. (23) does not follow. The symmetric treatment in Appendix E does not repair this: Eq. (E1) still contains the resonance frequency Ω and provides no frequency-independent prediction. The paper should either derive the frequency independence from the framework or explicitly state the main theorem as conditional on this additional assumption.","section":"IV, after Eq. (22); App. E"},{"comment":"The key step from \"classical gravity forces collapse\" to \"the dynamics must be diffusive\" is imported from ref. [84], which is authored by three of the present authors. The manuscript does not state the hypotheses or content of that theorem, although this step is load-bearing for the rest of the paper. Please include a self-contained statement of the theorem and its conditions, so that the reader can check that the collapse dynamics required by the no-signaling argument in Section II falls within its scope.","section":"II, ref. [84]"}],"minor_comments":[{"comment":"The step called \"the strongest condition\" is not a straightforward maximization of the left-hand side, because α also enters the right-hand side through the cosα and sinα terms; Eq. (20) is better described as the necessary condition obtained by choosing α=π/2.","section":"IV, Eq. (20)"},{"comment":"There are typos in the text: \"the the Hamiltonian\" and \"supposed to to equal\" should be corrected.","section":"III"},{"comment":"The text contains typos \"findinig\" and \"Wiener-Kinchine\" (should be Wiener-Khinchin).","section":"Appendix E"},{"comment":"The statement that the required parameters are within reach of near-term technology should be softened: the text itself notes a gap of four orders of magnitude in Q relative to demonstrated torsion pendulums, so the feasibility claim rests on extrapolation from loss scaling and on the Braginsky-Caves-Thorne estimate, not on demonstrated performance.","section":"V, Table I"},{"comment":"The notation fk(ĉ) is replaced by fk without further definition in Eq. (7); the notation should be harmonized for readability.","section":"III, Eqs. (5)-(7)"}],"recommendation":"major_revision","confidential_remarks":"The central quantitative claim is unsupported by the derivation as written, but the flaw is localized to the first-order expansion technique and to the unproved frequency-independence assumption; it is plausibly fixable within the scope of the paper. The reliance on ref. [84] for a load-bearing theorem may warrant editorial attention, since three of the present authors are also authors of that reference."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: the paper's central new number — the lower bound γ11 + m²ω²γ33 ≥ Gm²/(ħd³) — is not derived. The first-order PPT expansion in Eq. (18) is algebraically off. The vector z0=(a,-b,ia,ib) is not in the kernel of V0 + (i/2)ΛJΛ; with the ordering used, the kernel is p=-ix, so the correct null vector is (a,b,-ia,-ib) up to signs. On that true null space the first-order contribution of the K x1 x2 coupling vanishes: dV/dt from the interaction is zero because the initial product ground state has no cross-correlations and the Newtonian term is quadratic in x. So the first-order expansion cannot produce the K-dependent inequality that leads to Eq. (23). The bound may be true, but this paper does not prove it.\n\nWhat is genuinely good here: the paper makes the case, clearly and with proper citations, that classical local gravity must be diffusive even for classical probes, and that this suggests a much simpler experimental route than massive superpositions. The torsion-pendulum protocol and the reheating-rate method are concrete and well explained. The authors also honestly report the four-orders-of-magnitude gap between current Q/T and what is needed; the 'near-term' language in the abstract is doing a lot of work, but the main text is more careful.\n\nSoft spots beyond the derivation: the frequency-independence assumption on γ_ij is introduced as a 'reasonable expectation' rather than derived; even granted, it doesn't fix the expansion problem. The steps from Eq. (18) to Eq. (22) are too compressed to check without going through the algebra by hand, and the algebra turns out not to support the conclusion.\n\nWho this is for: people working on gravity-induced decoherence and tabletop quantum gravity tests. It is a serious contribution to the conversation, but the advertised bound is the load-bearing result and it doesn't hold up as written. The right move is to send it out for review with a clear request: redo the expansion at second order in the interaction and see what bound you actually get. If the K² order yields an inequality, the experimental requirements will likely change; if no clean bound emerges, the paper becomes a purely conceptual proposal. Either way, a good referee will earn their keep.\n\nRecommendation: send to peer review, expect major revision.","headline":"Clever idea, but the paper's advertised lower bound is not derived — the first-order PPT expansion has a wrong null vector and the Newtonian term vanishes at that order.","tokens_in":27166,"tokens_out":9894,"would_cite":false,"duration_ms":93275,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["04.60.-m","03.65.Yz","04.80.Cc"],"model":"deepseek-v4-flash","headline":"If gravity is classical — a local classical channel (LOCC) — it must diffuse the momentum of quantum matter by a minimum, calculable amount, and a millikelvin torsion pendulum could detect the effect without macroscopic quantum…","keywords":["classical gravity","LOCC","momentum diffusion","Lindblad master equation","gravitational decoherence","torsion pendulum","separability bound","non-unitary gravity"],"falsifier":"Run the proposed protocol — two 2.55 kg osmium masses on a torsion pendulum at $\\Omega/2\\pi = 10^{-4}$ Hz, cooled to 10 mK with quality factor $Q = 2\\times10^{10}$, feedback-cooled and then monitored in the dark with a near-quantum-limited detector — and measure the non-thermal phonon heating rate after thermal background subtraction. If the inferred diffusion combination $\\gamma_{11} + m^2\\omega^2\\gamma_{33}$ falls below $Gm^2/(\\hbar d^3)$ (equivalently, if the heating rate stays below $\\Gamma_G = \\pi\\omega_G^2/(12\\beta^3\\Omega)$), the paper's bound is violated and classical LOCC gravity is refuted. A second, targeted check: run the same masses at a different trap frequency; if the measured diffusion combination changes with $\\omega$ while $m$ and $d$ are fixed, the universality premise fails and only a setup-specific constraint would remain.","tokens_in":26136,"feed_emoji":"⚖️","tokens_out":20070,"duration_ms":183086,"temperature":0.7,"pith_summary":"This paper establishes a consequence of classical gravity: if gravity is a local classical channel (LOCC), it cannot act on quantum matter without also jiggling it. Avoiding faster-than-light signaling forces the classical interaction to collapse spatial superpositions, and a random, translation-invariant collapse is necessarily diffusive, so the jiggling is unavoidable even for ordinary, well-localized masses. For two harmonically trapped masses, the paper derives the minimum diffusion any such gravity must produce, $\\gamma_{11} + m^2\\omega^2\\gamma_{33} \\ge Gm^2/(\\hbar d^3)$, and shows that the effect appears as a slow reheating that a millikelvin torsion pendulum could in principle detect. The payoff is a table-top test of the quantum nature of gravity that requires no macroscopic superpositions and no quantum state control.","feed_headline":"If gravity is classical, it must jiggle matter—new bound sets the size","feed_subtitle":"A millikelvin torsion pendulum could detect the predicted shake without macroscopic quantum superpositions.","key_machinery":"The load-bearing object is the no-entanglement bound of Eq. (23), $\\gamma_{11} + m^2\\omega^2\\gamma_{33} \\ge Gm^2/(\\hbar d^3)$, obtained by requiring that the two oscillators' initial ground state passes the positivity-under-partial-transposition (PPT) criterion at all times. The master equation (11) is the second piece of machinery: a Lindblad form whose Hamiltonian carries the linearized Newtonian coupling $K\\hat{x}_1\\hat{x}_2$ and whose double-commutator terms, weighted by a real symmetric matrix $\\gamma_{ij}$, are the only way to add diffusion without spoiling the classical Newtonian limit. The third ingredient is the statistical argument tying classical gravity to collapse: the density-matrix map must be linear to prevent superluminal signaling, so the collapse is random, and translation-invariant random collapse is diffusive, which justifies the master-equation structure rather than an arbitrary noise model. The setup-independent form of the bound relies on treating the coefficients $\\gamma_{ij}$ as universal, depending on masses and distance but not on trap frequencies; the paper flags this as a reasonable expectation and discusses a symmetric-setup relaxation in its appendix.","core_discovery":"The central claim is that a classical, local gravitational interaction must come with a diffusive noise floor: the requirement that gravity never entangle two initially separated quantum systems, together with the no-signaling constraint, implies an unavoidable randomness in the gravitational coupling, and that randomness shows up as momentum diffusion in any probe, classical or quantum. The paper develops this through Feynman's 1957 thought experiment: classical gravity must collapse spatial superpositions, the collapse must be random, and the corresponding density-matrix map must be linear, which pins the dynamics to a specific master equation form rather than an ad hoc model. Within that master equation for two trapped masses with linearized Newtonian gravity, demanding that the ground state never entangle forces the diffusion coefficients to satisfy the lower bound of Eq. (23); the bound is necessary but not sufficient, so seeing the diffusion would not prove gravity is classical, while not seeing it at the required level would rule out classical gravity in the LOCC sense. The sharp experimental consequence is a minimum reheating rate for a cooled torsion pendulum, quantified by $\\Gamma_G = \\pi\\omega_G^2/(12\\beta^3\\Omega)$ with $\\omega_G = \\sqrt{G\\rho}$, which near-term technology could in principle resolve.","pith_inferences":["The argument's logic is not specific to gravity: any classical local channel mediating a force between quantum systems would need a similar noise floor, so the inequality's structure suggests analogous diffusion bounds for other hypothetical classical forces, each with its own coupling constant in place of $G$.","The universality premise is directly testable: repeat the proposed measurement at two different trap frequencies with the same masses and separation; if the inferred combination $\\gamma_{11} + m^2\\omega^2\\gamma_{33}$ changes with $\\omega$, the setup-independent bound fails and only weaker, setup-specific constraints survive.","The quiet-environment requirement might be easier to meet in space than on Earth: the paper notes that LISA Pathfinder already demonstrated acceleration noise below $10^{-15}\\,g/\\sqrt{\\mathrm{Hz}}$, which would sidestep the millikelvin torsion-fiber dissipation challenge altogether.","A single high-Q torsion pendulum of this kind would simultaneously probe the gravitational diffusion bound and the parameter space of spontaneous collapse models, since both predict the same double-commutator heating signature."],"forward_implications":["A null result — a clean experiment that limits the diffusion below Eq. (23) — would rule out all classical LOCC models of gravity without ever preparing a macroscopic superposition.","Every existing hybrid classical-quantum gravity model (collapse models, feedback-based models, the postquantum theory) reduces to the same master equation structure with different diffusion coefficients, so the bound constrains all of them jointly.","The proposed torsion-pendulum protocol is quantitatively specified: $\\Omega/2\\pi = 10^{-4}$ Hz, osmium masses of 2.55 kg at near-contact separation, $T = 10$ mK, $Q \\approx 2\\times10^{10}$, a near-quantum-limited readout, and about two days of integration to resolve 1% of the thermal noise.","Current torsion pendulums sit roughly four orders of magnitude below the required damping time, but the millikelvin regime was already estimated in 1977 to allow $Q \\sim 10^{10}$, and no dedicated attempt has been made since.","Detecting the diffusion would constitute the first observed departure from the unitary Schr\\\"odinger evolution under the Newtonian potential, marking a gravitational effect that is real but non-quantum."],"supporting_citations":[{"why":"The entanglement-witness protocol whose need for macroscopic quantum superpositions this work's diffusion test is designed to bypass.","marker":"[26]"},{"why":"The companion argument that classical gravity as a LOCC cannot entangle two initially separated systems, the premise this paper extends to diffusion.","marker":"[27]"},{"why":"Supplies the key theorem that any translation-invariant collapse dynamics must be diffusive, turning the no-signaling argument into a momentum-diffusion prediction.","marker":"[84]"},{"why":"Establishes that the density-matrix map must be linear to prevent superluminal signaling, which forces the collapse to be random.","marker":"[85]"},{"why":"The feedback-based classical-channel model of gravity whose two-particle master equation this work generalizes to the full $\\gamma$-matrix form.","marker":"[89]"},{"why":"The original positivity-under-partial-transposition separability criterion used to constrain the diffusive dynamics.","marker":"[100]"},{"why":"The continuous-variable formulation of PPT as a partial phase-space reflection, which converts the criterion into the matrix inequality (16).","marker":"[102]"},{"why":"An earlier noise inequality for non-entangling Lindblad dynamics, the direct precedent for the lower bound in Eqs. (21)-(23).","marker":"[103]"},{"why":"The 1977 estimate that millikelvin torsion pendulums can reach $Q \\approx 10^{10}$ and damping times beyond $10^{13}$ s, the feasibility anchor of the proposed experiment.","marker":"[111]"}],"fun_headline_variants":["Classical gravity forces diffusion—new bound enables a near-term test","No huge superpositions: detect classical gravity's jiggle with a pendulum","Gravity if classical must diffuse—torsion pendulum can catch it","New bound: classical gravity's noise is measurable without quantum states","A millikelvin pendulum could spot the shake classical gravity would cause"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the gravitational diffusion coefficients are universal — independent of the trap frequency, fixed only by the masses and their separation — which the paper calls a 'reasonable expectation' but does not derive; if classical gravity diffused matter in a frequency-dependent way, only a weaker, setup-specific bound would follow, and the experiment would not test gravity itself.","fun_headline_variants_meta":{"raw":{"variants":["Classical gravity forces diffusion—new bound enables a near-term test","No huge superpositions: detect classical gravity's jiggle with a pendulum","Gravity if classical must diffuse—torsion pendulum can catch it","New bound: classical gravity's noise is measurable without quantum states","A millikelvin pendulum could spot the shake classical gravity would cause"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000283,"raw_usage":{"total_tokens":1679,"prompt_tokens":957,"completion_tokens":722,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":573,"completion_tokens_details":{"reasoning_tokens":629}},"tokens_in":573,"tokens_out":722,"duration_ms":8174,"temperature":1.0,"reasoning_tokens":629,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T16:32:01.429350+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the proposed protocol — two 2.55 kg osmium masses on a torsion pendulum at $\\Omega/2\\pi = 10^{-4}$ Hz, cooled to 10 mK with quality factor $Q = 2\\times10^{10}$, feedback-cooled and then monitored in the dark with a near-quantum-limited detector — and measure the non-thermal phonon heating rate after thermal background subtraction. If the inferred diffusion combination $\\gamma_{11} + m^2\\omega^2\\gamma_{33}$ falls below $Gm^2/(\\hbar d^3)$ (equivalently, if the heating rate stays below $\\Gamma_G = \\pi\\omega_G^2/(12\\beta^3\\Omega)$), the paper's bound is violated and classical LOCC gravity is refuted. A second, targeted check: run the same masses at a different trap frequency; if the measured diffusion combination changes with $\\omega$ while $m$ and $d$ are fixed, the universality premise fails and only a setup-specific constraint would remain.","supporting_citations":[{"cited_title":"Search for spontaneous radiation from wave func- tion collapse in the majorana demonstrator","cited_arxiv_id":null,"evidence_quote":"Supplies the key theorem that any translation-invariant collapse dynamics must be diffusive, turning the no-signaling argument into a momentum-diffusion prediction."},{"cited_title":"Present status and future challenges of non-interferometric tests of collapse models","cited_arxiv_id":null,"evidence_quote":"Establishes that the density-matrix map must be linear to prevent superluminal signaling, which forces the collapse to be random."},{"cited_title":"Testing quantum mechanics","cited_arxiv_id":null,"evidence_quote":"The feedback-based classical-channel model of gravity whose two-particle master equation this work generalizes to the full $\\gamma$-matrix form."},{"cited_title":"Linear-friction many-body equation for dissipative spontaneous wave-function collapse","cited_arxiv_id":null,"evidence_quote":"The original positivity-under-partial-transposition separability criterion used to constrain the diffusive dynamics."},{"cited_title":"Separability criterion for density matrices","cited_arxiv_id":null,"evidence_quote":"The continuous-variable formulation of PPT as a partial phase-space reflection, which converts the criterion into the matrix inequality (16)."},{"cited_title":"A measurement of g with a cryogenic torsion pen- dulum","cited_arxiv_id":null,"evidence_quote":"The 1977 estimate that millikelvin torsion pendulums can reach $Q \\approx 10^{10}$ and damping times beyond $10^{13}$ s, the feasibility anchor of the proposed experiment."}],"review_version":1}