Dark Matter or Modified Gravity? The Universe's Thermal Bath Might Decide

Emergent Gravity and MOND — Quantum-Geometric Correspondence series

The Dark Matter Problem

For nearly a century, astronomers have faced an embarrassing problem: galaxies do not rotate the way they should.

When you measure how fast stars orbit the centers of spiral galaxies, you find something strange. Stars in the outer regions move far too quickly. According to Newton's laws and the visible matter we can see, those outer stars should be moving slowly---the gravitational pull from the galaxy's center falls off with distance, so distant stars should lazily drift along their orbits.

Instead, they race around at nearly the same speed as stars much closer to the center. The rotation curves are flat where they should be declining.

The standard explanation is dark matter: invisible stuff that outweighs ordinary matter five to one, forming vast halos around galaxies that provide the extra gravitational pull needed to keep those outer stars moving fast. Dark matter has become a cornerstone of modern cosmology, essential for explaining not just galaxy rotation but also gravitational lensing, cosmic structure formation, and the pattern of fluctuations in the cosmic microwave background.

But there is a problem. Despite forty years of searching, no one has directly detected a dark matter particle. Underground detectors, particle accelerators, space telescopes---all have come up empty. The particles that should constitute this cosmic scaffolding remain stubbornly invisible not just to telescopes but to every experiment designed to catch them.

What if the answer is not new matter, but new physics?

MOND: A Pattern That Refuses to Die

In 1983, Israeli physicist Mordehai Milgrom proposed a radical alternative: maybe Newton's laws break down at very low accelerations.

His theory, called Modified Newtonian Dynamics (MOND), introduced a single new constant of nature: a characteristic acceleration scale $a_0 \approx 1.2 \times 10^{-10}$ meters per second squared. This is an incredibly small acceleration---about a hundred billion times weaker than Earth's surface gravity.

Milgrom's proposal was simple: when gravitational accelerations fall below $a_0$, the effective gravitational force becomes stronger than Newton predicts. The modification is precisely calibrated to produce flat rotation curves.

The physics community largely dismissed MOND as an ad hoc fix---a mathematical trick rather than real physics. But MOND has a disturbing track record of successful predictions:

• Flat rotation curves: Explained by construction, but MOND also predicts their detailed shapes
• The Tully-Fisher relation: The observed relationship between galaxy luminosity and rotation speed ($v^4 \propto M$) emerges automatically from MOND, while dark matter models must tune parameters to match it
• Low surface brightness galaxies: MOND predicted their rotation curves before they were measured
• The acceleration scale itself: $a_0 \approx cH_0$, suspiciously close to the Hubble expansion rate times the speed of light

That last point is particularly haunting. Why should a parameter governing galaxy dynamics have anything to do with cosmology? In dark matter models, this is pure coincidence. In MOND, it hints at something deeper.

The Thermodynamics of Cosmic Horizons

Here is where our story takes an unexpected turn.

We live in an accelerating universe. The expansion is speeding up, driven by what we call dark energy. This acceleration creates a fundamental boundary in spacetime: the cosmological horizon. Objects beyond this horizon are receding from us faster than light and can never send us signals. We are surrounded by an information barrier.

In 1977, Gary Gibbons and Stephen Hawking showed that this cosmological horizon has thermodynamic properties, just like a black hole horizon. It has a temperature:

$T_{dS} = \frac{\hbar H_0}{2\pi k_B}$

This is the Gibbons-Hawking temperature of de Sitter space---roughly $10^{-30}$ Kelvin, unimaginably cold but not zero. Our universe is bathed in a thermal background set by its own expansion.

Horizons also have entropy. The famous Bekenstein-Hawking formula tells us that black hole entropy is proportional to horizon area. But there is another contribution: when you have a thermal bath filling a region of space, you also get volume entropy---entropy proportional to the volume rather than the area.

For most purposes, this volume entropy is negligible. But something interesting happens at very low accelerations.

When Volume Entropy Takes Over

The key principle is what we call entanglement equilibrium: in any local region, the total generalized entropy (area entropy plus volume entropy) should be stationary under small variations. This is a thermodynamic statement about how gravity and quantum information balance.

In strong gravitational fields, the area term dominates, and entanglement equilibrium gives you ordinary general relativity. Einstein's equations emerge as a thermodynamic identity.

But at very low accelerations---when $g \lesssim a_0$---we postulate that something changes. The volume entropy contribution becomes comparable to the area term, local entanglement equilibrium fails, and the effective gravitational dynamics shift. It is worth being clear here: this failure of local equilibrium at low accelerations is a physical hypothesis, not a derived result. It is the central assumption of the framework, well motivated but not proven from first principles.

The characteristic scale where this transition is taken to occur is:

$a_0 = \frac{c H_0}{2\pi} \approx 1.08 \times 10^{-10} \text{ m/s}^2$

This scale is not a free parameter. It follows from the Gibbons-Hawking temperature of de Sitter space and the requirement of thermodynamic consistency, with no adjustable knobs. The value matches the observed MOND acceleration scale within 10%.

Let that sink in: the mysterious MOND acceleration, which Milgrom introduced as an empirical fit to galaxy data, emerges from the thermodynamic properties of cosmic horizons with no adjustable parameters.

Where Does Einstein Still Rule?

If gravity is modified at low accelerations, why do we not see this in the solar system? Why do spacecraft trajectories follow Newtonian predictions so precisely?

The answer is regime separation. The relevant parameter is $\epsilon = g_N/a_0$, the ratio of Newtonian gravitational acceleration to the critical scale.

• Solar system: At Earth's surface, $g \approx 10$ m/s$^2$, so $\epsilon \approx 10^{11}$. We are deep in the Newtonian regime, with any MOND correction suppressed to the level of $10^{-11}$.
• Binary pulsars: Orbital accelerations give $\epsilon \approx 10^{13}$. General relativity is exact; gravitational wave predictions are confirmed to extraordinary precision.
• Galaxy outskirts: At 50 kpc from the Milky Way center, $\epsilon \approx 1$. This is exactly where MOND effects become significant.

The theory naturally explains why modifications appear only where they are observed: in the outer regions of galaxies, in dwarf galaxies, in low surface brightness systems---everywhere that accelerations drop below the cosmic threshold.

From Phenomenon to Field Theory

A scaling law is not a theory. To make contact with modern physics, we need a covariant field theory that reduces to MOND in the appropriate limit while remaining consistent with general relativity where it works.

We sketch such a framework, which we call Entanglement-Elastic Gravity (EEG). It is best understood as a schematic proposal rather than a finished theory. The idea is that deviations from entanglement equilibrium create a kind of "elastic strain" in spacetime---a displacement field that encodes how far the local entropy balance is from its equilibrium configuration.

The EEG action adds this elastic displacement field to the standard Einstein-Hilbert action. The version written down here, however, uses a linearized quadratic action---a caricature valid near the high-acceleration regime. Honesty requires noting that this linearized action does not by itself reproduce the modified Poisson equation of MOND: recovering the deep-MOND behavior $g \propto \sqrt{g_N}$ requires promoting the elastic field to a genuinely dynamical one with a nonlinear kinetic functional, of the Bekenstein-Milgrom (AQUAL) type, which we leave to future work. What the framework does fix, independent of that completion, are the acceleration scale $a_0 = cH_0/(2\pi)$ and the elastic modulus $\kappa = c^4/(8\pi G a_0)$.

EEG is designed to satisfy two requirements that any successful completion must preserve:

• Ghost-free: No negative-energy degrees of freedom that would make the theory unstable
• Speed-of-light preserving: Gravitational waves should travel at exactly $c$, consistent with LIGO/Virgo observations

These are design constraints on a future nonlinear completion, not yet established properties of a finished theory.

Testable Predictions

The framework points to several signatures that would distinguish it from both standard dark matter and other modified gravity proposals. Two of these (the lensing slip and the growth-rate suppression) are at this stage qualitative expectations rather than quantitatively derived numbers---the magnitudes below are order-of-magnitude targets for future surveys, not firm predictions.

1. Weak lensing slip: In general relativity, the two gravitational potentials $\Phi$ and $\Psi$ that affect light and matter are equal. The framework points to a slip between them:

$\frac{|\Phi - \Psi|}{|\Phi|} \sim 15\%$

at radii beyond 50 kpc. If borne out at this level, it would be measurable with weak lensing surveys.

2. Tully-Fisher relation: In the deep-MOND regime the theory gives $v^4 = GMa_0$, with the same $a_0$ that comes from cosmology. No tuning required. This is a parameter-free prediction already consistent with the observed relation.

3. Growth rate suppression: Structure formation in the universe should be slightly suppressed compared to $\Lambda$CDM predictions, plausibly at the level

$\Delta f\sigma_8 \sim -0.03$

at $z \approx 0.8$. The Euclid and DESI surveys are the natural place to test for an effect of this kind.

4. No dark matter detection: If the theory is correct, direct detection experiments will continue to find nothing, because there is nothing to find.

What This Means

If this framework is correct, several profound conclusions follow.

First, dark matter---at least as an explanation for galaxy dynamics---may not exist. The phenomena attributed to dark matter would instead reflect the thermodynamic properties of our universe's cosmological horizon. We would not need new particles, just a deeper understanding of how gravity, thermodynamics, and information connect.

Second, MOND would cease to be an ad hoc empirical formula and become a derived consequence of fundamental physics. The "unreasonable effectiveness" of MOND in fitting galaxy data would have an explanation: it captures the low-acceleration regime where cosmic thermodynamics modifies gravitational dynamics.

Third, the cosmic coincidence $a_0 \approx cH_0$ would be explained. This is not a coincidence at all; it is a direct consequence of the de Sitter horizon temperature setting the scale where entanglement equilibrium breaks down.

Honest Uncertainties

Scientific integrity requires stating what we do not know.

The framework does not explain everything attributed to dark matter. Galaxy cluster dynamics, the cosmic microwave background power spectrum, and certain gravitational lensing observations all present challenges. Some additional ingredient---perhaps sterile neutrinos, perhaps something else---may still be needed on large scales.

The covariant completion (EEG) is a schematic proposal, not a derived result. The linearized version we wrote down does not yet recover MOND in the Newtonian limit---that step requires a nonlinear (AQUAL-type) kinetic functional we have not constructed---and even granting it, we have not proven the completion is unique or that other completions are impossible.

The transition regime where $\epsilon \sim 1$ requires careful treatment that may reveal complications we have not anticipated.

Despite these caveats, the core observation stands: the MOND acceleration scale emerges from cosmic thermodynamics with no free parameters. This demands explanation regardless of the ultimate fate of any specific theoretical framework.

This is the emergent-gravity paper of the Quantum-Geometric Correspondence series, exploring how MOND phenomenology emerges from the thermodynamics of de Sitter space.