Physicist review: Verlinde-Erik, "On the Origin of Gravity and the Laws of Newton" (arXiv:1001.0785)

Reviewer: PhD physicist (Step 2). Sources read: facts.json, claims.md, deep/gravity-paper-42e8a5.md (read in full), sites.md "Leads for review", research/ notes (arxiv, nature, others skimmed). research/files/ is empty (the Investigator downloaded nothing), so every Investigator item is known to me only through the Investigator's notes plus the three abstracts I re-fetched. Tags: [read] document/notes, [web] URL, [memory] with confidence, [calc] deep reader's checked algebra.

No source gap: the full paper text was extracted. Conclusions about later literature rest on [web] abstracts, not on the full papers.

Corrections to Step 1

1. facts.json says Sections 1-3 and 6 rest on digest only. True, so my reading of Section 3 (the entropy postulate) is via the deep read's quotations (pp.7-11) [read]. I did not re-open page images; the equations I rely on are the deep read's checked ones. Flag: the central postulate (3.5)-(3.6) is the one place not independently checked by page image.
2. The deep read says the R^2/d^2 check "verifies" (3.6) as d -> R. Correct, but it notes the local flat-patch integral gives half the 2 pi. That is a real limitation of the postulate (it works only for a closed screen), and claims.md does not carry it. Add it to claim 2/3.
3. The deep read describes (5.27) as a sign slip. Agreed on the sign (a static hovering observer accelerates outward), but the Investigator's lead says Verlinde 2026 argues the sign is correct because the entropy is invariant (an "adiabatic reaction force"). That is the inventor's later reading, not a correction of the 2010 text; I keep both.
4. facts.json says "[memory: medium] later tests include neutron interferometry". The accurate statement [web] is ultra-cold neutron bound states in Earth's field (Kobakhidze, https://arxiv.org/abs/1009.5414), not interferometry. Correction made.
5. sites.md lists "Mäkelä 2010 (arXiv:1001.3808): not opened". I opened it [web https://arxiv.org/abs/1001.3808]: title "Notes Concerning 'On the Origin of Gravity and the Laws of Newton' by E. Verlinde". Mäkelä's abstract states that equations Verlinde postulates as a starting point "are actually consequences of a specific microscopic model of spacetime, which has been published earlier" (his own). This is a priority/precedence claim by an interested party (Mäkelä is the author of the earlier model); I did not verify his model. It is kept in the record.
6. Visser (1108.5240) abstract re-checked [web]: says Newtonian gravity being conservative "places significant constraints on the form of the entropy and temperature functions", causing problems for Verlinde's proposal but not Jacobson's or Padmanabhan's. Matches the Investigator.

What the mainstream view is, fairly stated

• Newtonian gravity and GR are well tested: lab torsion balances, planetary ephemerides, lunar laser ranging, binary pulsars, LIGO. Any reformulation must reproduce them. The paper does (by construction).
• Thermodynamic gravity has respectable precedent: Bekenstein-Hawking black-hole entropy S = A c^3/(4 G hbar) k_B, Hawking temperature, Unruh effect, Jacobson 1995 (Einstein equation as an equation of state, from dQ = T dS on local horizons) [memory: high], Padmanabhan [web via Investigator]. Mainstream treats this as a hint that gravity has a statistical origin, not as an established derivation. The Unruh effect itself has never been directly measured [memory: high].
• The mainstream complaint about Verlinde's paper is not that the formulas are wrong but that they are built from their results (see circularity below), and that the "entropic force" reading faces specific problems (Visser, Wang-Braunstein, neutron states).

Claim-by-claim

Claim 1: Gravity is an entropic force, F = T dS/dx

a. Mathematics: Internally consistent in the pieces given (F dx = T dS; algebra verified [calc]). Sign conventions are loose in (4.25) and (5.27)/(5.30), magnitudes right. Visser [web] shows a deeper issue: an entropic force that must also be conservative (as Newtonian gravity is) tightly constrains S(x) and T(x). A single-heat-bath polymer analogy does not satisfy this for general potentials; he needs multiple baths. The paper's temperature varies over the screen and "equilibrium" is not assumed (p.13), so the polymer analogy (one bath, one T) is weaker than presented. b. Experiment: A genuine entropic force is a statistical, dissipative effect; it should show noise, decoherence and temperature dependence. Gravity shows none at the levels tested. Colella-Overhauser-Werner (1975) saw a coherent quantum phase shift of neutrons in Earth's field [memory: high]; quantised bounce states of ultra-cold neutrons (Nesvizhevsky 2002) show gravity acting as a clean potential in the Schrodinger equation [memory: high]. Kobakhidze argues these "disprove" the entropic origin [web https://arxiv.org/abs/1009.5414]. The counter-view is Chaichian-Oksanen-Tureanu: EG need not contradict bound states, since EG is Newtonian there, though spontaneous decays of excited states would need extra ingredients [web https://arxiv.org/abs/1104.4650, via Investigator]. The inventor's later reply (2026) is that the slow degrees of freedom stay coherent (Born-Oppenheimer) so the force is a reaction force "acting without increasing the entropy" [read: Investigator note, research/2026-10-04-worldscientific.md; not independently checked by me]. That reply makes the "entropic" label closer to an adiabatic elastic force, which is a different claim from the polymer-in-a-heat-bath picture of 2010. c. Tested? Not as a distinct theory in 2010: it predicts the same forces as Newton. The only direct discriminators are the neutron-state decays and force fluctuations (claim 6), neither performed as a decisive test. No replication in the sense of a test confirming a difference. d. If true: gravity would be derived from a microscopic theory of the holographic degrees of freedom, with G a derived quantity of the bit density. Quantum gravity would not be "quantising the metric". But the paper gives no microscopic theory. Verdict: Split. Strong form (gravity is literally a thermodynamic entropy-driven force, with decoherence/noise as in a polymer): Never properly tested in the form stated and under strain from Visser and the neutron argument (weak-to-contested; I regard it as not supported by existing data but not cleanly tested). Weak form (gravity's equations have a thermodynamic structure): Plausible, untested as an interpretation, with the Jacobson-type result as its respectable core.

Claim 2: Inertia is entropic (F = ma from dS = 2 pi k_B (mc/hbar) dx and Unruh T)

a. Algebra: F = T dS/dx with k_B T = hbar a/(2 pi c) gives F = ma exactly [calc]. But the 2 pi in the postulate "is picked precisely" to give F = ma (p.8) and the Unruh form is reversed ("temperature required to cause an acceleration", p.8), then declared unnecessary ("forget about the Unruh law", p.8) yet reused for T in 4.1 and 5.1. So the "derivation" is a consistency identity: input constant chosen to output the answer. The postulate also holds only for a closed screen (flat-patch gives half the 2 pi) [calc]. b. Experiment: F = ma is tested to high precision. The entropic reading gives no different number. The Unruh temperature for accelerations achievable in the lab is tiny (T = hbar a/(2 pi k_B c); for a = 9.8 m/s^2, 4e-20 K [calc]) and unobserved directly [memory: high]. c. Tested: no distinguishing test. d. If true: inertia would be a property of information bookkeeping, tied to the Higgs-independent origin of mass in no obvious way; the paper does not say how rest mass m enters other than via the Compton wavelength. Verdict: Not testable as stated (as an interpretation it makes no measurable difference from F = ma); with respect to the claim of "derivation", Internally inconsistent in one respect only in logic, not equations: the Unruh law is both discarded and used, and the constant is fitted to the answer. I assign Not testable as stated, and I list the circularity as a flaw to the claim of derivation.

Claim 3: Newton's law and the Poisson equation follow from holography + equipartition

a. Algebra verified: (1/2)k_B T dN integrates to Gauss's law (4.20) [calc]; Earth numbers: T = 3.98e-20 K, N = 1.95e84 bits, (1/2)N k_B T = Mc^2 [calc]. Poisson follows from Gauss's law for arbitrary surfaces by the divergence theorem [memory: high]; the paper imposes it for equipotential surfaces and by a "bit of mass pushed across" argument (p.14), a gap the deep read notes. The factor N = Ac^3/(G hbar) "can simply be regarded as the definition of G" (p.8). The circularity ledger [read: deep read section 9] shows the inputs: Newton potential as a "device to describe the local acceleration", E = Mc^2 and M as enclosed mass, the Unruh T. The output is exactly Newton. The author states: "It had to work ... the laws of Newton have been ingredients in the steps that lead to black hole thermodynamics and the holographic principle. In a sense we have reversed these arguments." (p.9) [read]. That is honest, and it means no independent confirmation of the Newton form is obtained. Worked check I can add by hand: bits on a sphere of radius R: N = 4 pi R^2 c^3/(G hbar). With T = hbar g/(2 pi k_B c) and E = (1/2) N k_B T = (1/2)(4 pi R^2 c^3/(G hbar))(hbar g/(2 pi c)) = R^2 c^2 g/G. Setting equal to Mc^2 gives g = GM/R^2. G cancels against its own definition, which is why the result cannot test G. b. Consistency with experiment: trivially yes for Newton. Equipartition "in general holds only for free systems" and is the paper's own "least obvious assumption" (p.10). The bits' implied entropy appears to violate the Bekenstein bound (p.28) [read]. c. Tested: not independently. Wang-Braunstein 2018 [web https://www.nature.com/articles/s41467-018-05433-9, via Investigator]: the first law does not hold on ordinary surfaces away from horizons except under spherical symmetry, "undermining the key thermodynamic assumption" of the programme. A reader comment says this assumes an area law whereas the 2016 model has a volume law; that comment is one commenter's view, kept on record. Their result is a theoretical analysis, not an experiment. Visser as above. d. If true: G would not be fundamental; the area-law degree-of-freedom count would be the origin of Newton's constant. No change to predictions. Verdict: Split. Reproduction of Newton/Poisson as algebra: Established (as a consistency identity). Claim that this explains or derives gravity: Internally inconsistent in the weak sense of circular (outputs used as inputs; G defined by the count; 2 pi fitted), so I cannot award "derived". Closest scale entry for the derivation claim: Not testable as stated (it makes no different prediction), with Tested, not supported applying only to the general-potential generalisation under Visser/Wang-Braunstein (theory, not experiment).

Claim 4: Relativistic version: F = -m e^phi grad phi, Komar mass, Einstein equations

a. (5.30), (5.31), (5.34), (5.39) follow by algebra [calc]. The Einstein-equation step is "a sketch" by the author's own word (p.20), stress-energy side "presumably fixed" by comparison; approximate Killing vectors in small regions leave error terms undiscussed; passage from a contracted component to the tensor equation is one paragraph. Section 5.3's (5.38)-(5.39) is a generic first-law identity that "did not really use ... the Einstein equations" (p.22), so by itself it cannot single out GR. The cosmological-constant term is not mentioned [memory: medium; consistent with Jacobson's well-known freedom for a Lambda term, which I did not re-verify on the web]. The inputs (static spacetime, Killing vector, Komar mass, Wald pp.288-290) are GR itself. b. Experiment: reproduces GR where GR is tested, nothing else. The paper's own 2026-restated sign point is above. c. Tested: no. Closest precursors: Jacobson 1995, Padmanabhan, whose results are accepted as suggestive. Wang-Braunstein targets exactly this (screens off horizons). d. If true: Einstein equations would be equations of state; GR a hydrodynamic-like limit. Verdict: Plausible, untested (as a heuristic thermodynamic reading, with Jacobson's derivation as the sounder version); the specific Verlinde route is a sketch, and the Wang-Braunstein result is a theoretical strike against its key assumption.

Claim 5: Space, geometry and gravity are emergent; end of gravity as fundamental force; redshift is a consequence of entropy gradients

a. Not mathematics but program. AdS/CFT is a precise duality in a specific negatively-curved setting [memory: high]; transferring "bulk emerges from boundary" to flat space and our universe is an extrapolation. The paper concedes the argument is "rather heuristic" (p.22-23) and the microscopic starting point is not given (p.19). b. Experiment: no data either way at 2010. The 2016 follow-up (Emergent Gravity and the Dark Universe, arXiv:1611.02269 [web via Investigator]) did produce a new low-acceleration prediction using a_0 = cH_0 and a volume law, which is a different model from the 2010 paper (area law only). Tests of that model, per the Investigator: Brouwer et al. 2017 KiDS-GAMA lensing "good agreement" with no free parameters (authors say "only a first step") [web https://arxiv.org/abs/1612.03034]; KiDS-1000 2021 (Verlinde a co-author, so an interested party): agreement with EG and MOND predictions, but a difference of at least 6 sigma between early- and late-type galaxies that modified gravity independent of galaxy properties "would be unable to explain" [web https://arxiv.org/abs/2106.11677]; Pardo (isolated dwarfs): Tully-Fisher reproduced, but velocities "severely underpredicted" above ~100 km/s [web https://arxiv.org/abs/1706.00785]; Nieuwenhuizen (Abell 1689) said EG fails without eV neutrinos [web, via Wikipedia, primary not opened]. I did not open these; they are quoted from the Investigator and belong to the 2016 model, not the 2010 paper. They are mixed results and must be reported as mixed: they neither confirm nor refute the 2010 claim. c. Tested: the 2010 claim as such, no. The 2016 offshoot, yes, with mixed outcomes, partly by the author's own group. d. If true: GR and quantum theory would both be effective theories; the quantisation of gravity would be moot; "dark gravity" possibly replaces dark matter, in conflict with the CMB and structure-formation successes of LambdaCDM [memory: high], unless the extension handles them (not shown). Verdict: Plausible, untested for the 2010 form; for the 2016 extension Tested with mixed results (not supported by dwarf and early/late-type data, supported by lensing at the first-step level), and the neutral controlled test is not on record.

Claim 6: Gravitational force fluctuations, larger for weak fields and small bodies

a. Stated qualitatively only; the author says better theory is needed (p.28). b. No known measurement. c. Later work: Verlinde-Zurek [web https://arxiv.org/abs/1902.08207, via Investigator] built a specific model of "geontropic" fluctuations; the GQuEST interferometer (Vermeulen et al., PRX 15, 011034 (2025) [web https://arxiv.org/abs/2404.07524]) is a design paper; I found no results. Carney et al. [web https://arxiv.org/abs/2409.03894] say standard graviton effective field theory predicts Delta L of order 1e-35 m, so a large signal would mean an effective-field-theory breakdown. Zurek is a co-author of the model: interested party. d. If true: EFT of gravity would fail at low energy, an enormous change. Verdict: Never properly tested (the experiment is being built; no dismissal without test). As stated in 2010: Not testable as stated.

Investigator findings kept in the record (whatever the verdict)

• Inventor's own: 2026 IJMPA note (adiabatic reaction force, Utrecht colloquium Dec 2009, inertia as his addition to Jacobson); 2016 EG paper; Verlinde-Zurek 2019-2022; Strings 2011 talk (not read).
• Support by others: Brouwer 2017 (lensing, good agreement); Brouwer 2021 (agreement plus 6-sigma tension); Chaichian et al. (neutron states compatible); Pardo (Tully-Fisher OK, dwarfs poor); GQuEST design.
• Related/critique: Jacobson 1995; Padmanabhan 2010; Cai-Cao-Ohta, Easson-Frampton-Smoot, Smolin follow-ups; Visser 2011 (problems unless multiple heat baths); Wang-Braunstein 2018; Kobakhidze 2011 (neutrons); Nieuwenhuizen 2017; Carney 2026; Mäkelä 2010 (precedence claim, interested party); Bianconi 2025 and Parker-Jeynes as other entropic-gravity proposals.
• Historical: first announced 12 Dec 2009 in de Volkskrant (not read, blocked); NYT/New Scientist/Wired coverage 2010 (Strominger quoted via Wikipedia as reactions from "can't be right" to "right and we already knew it"); Spinoza Prize 2011; JHEP 04 (2011) 029. The record does not go silent: this is a live programme, so no absence-of-evidence analysis is needed. The Investigator did not read the Volkskrant piece, Not Even Wrong, NYT, or the Strings 2011 slides; contemporary reception is thin in the record, and this review adds no claim about it.
• Not verified by me: Wang-Braunstein, Brouwer, Pardo, Nieuwenhuizen details beyond the Investigator's notes. Hossenfelder's critique (title recalled from memory by the Investigator, low confidence) is not cited.

Summary of verdicts

1. Entropic gravity: strong form Never properly tested, contested by Visser and neutron argument; weak (thermodynamic structure) form Plausible, untested.
2. Entropic inertia: Not testable as stated; "derivation" is fitted (2 pi chosen to give F = ma).
3. Newton/Poisson from holography: Established as algebra; as a derivation circular (G defined by bit count, Unruh T and potential as inputs), no independent test.
4. Relativistic/Einstein: Plausible, untested (author's own "sketch"); Wang-Braunstein challenges the key assumption.
5. Emergent spacetime: Plausible, untested (2010); 2016 extension tested with mixed results.
6. Force fluctuations: Never properly tested; not testable as stated in 2010.