Entropic Gravity Concepts

Entropic gravity—the idea that gravity is not a fundamental force but an emergent, thermodynamic phenomenon—has fascinated physicists, astronomers, and even philosophers for the past two decades. It sits at the crossroads of quantum theory, general relativity, and statistical mechanics, promising a unified language for phenomena that range from black‑hole horizons to the sprawling filaments of the cosmic web.

For the Apiary community, which cares deeply about the health of pollinator ecosystems and the stewardship of autonomous AI agents, entropic gravity offers a compelling analogy: just as a beehive’s collective behavior arises from simple, local rules, the large‑scale structure of the universe may spring from microscopic information dynamics. Moreover, the same mathematical tools that describe entropy flow in a hive can be repurposed to model the “self‑governance” of AI agents that learn, adapt, and coordinate without central control.

In this pillar article we take a hard look at Erik Verlinde’s emergent gravity proposal (2016), evaluate its quantitative predictions against modern gravitational‑lensing and galaxy‑cluster observations, and explore where the theory stands within the broader landscape of cosmology. Along the way we draw honest bridges to bee biology and AI governance, showing how the same concepts of information, temperature, and entropy echo across scales—from the honeycomb to the cosmos.

1. Foundations: Thermodynamics, Entropy, and Gravity

The link between thermodynamics and gravity was first hinted at by Jacob Bekenstein (1972) and Stephen Hawking (1974), who showed that black‑hole horizons carry an entropy

\[ S_{\rm BH}= \frac{k_{\rm B}c^{3}}{4 \hbar G}A \approx 1.07\times10^{77}\,{\rm k_{B}}\, \Bigl(\frac{M}{M_\odot}\Bigr)^{2}, \]

where \(A\) is the horizon area and \(M\) the black‑hole mass. The corresponding temperature

\[ T_{\rm H}= \frac{\hbar c^{3}}{8\pi k_{\rm B} G M} \]

implies that black holes radiate (Hawking radiation) and, crucially, that gravity knows about entropy.

In statistical mechanics, entropy quantifies the number of microscopic configurations compatible with a macroscopic state. The second law—entropy never decreases in an isolated system—drives the direction of time and governs the flow of heat. If gravity can be expressed as a response to entropy gradients, then spacetime itself may be a macroscopic description of an underlying information‑theoretic substrate.

This idea resonates with the way a bee colony regulates temperature: individual bees exchange heat through their bodies, collectively maintaining a stable hive temperature (≈ 35 °C) despite external fluctuations. The colony’s “thermostat” is not a central controller but an emergent property of local interactions. Similarly, emergent gravity posits that the curvature of spacetime is a statistical response to the distribution of microscopic degrees of freedom—often called bits—that live on holographic screens surrounding matter.

2. Jacobson’s 1995 Derivation: Einstein Equations as an Equation of State

Ted Jacobson (1995) offered the first rigorous connection between thermodynamics and Einstein’s field equations. He considered a local Rindler horizon—a patch of a null surface perceived by an accelerated observer—and imposed the Clausius relation

\[ \delta Q = T\,\delta S, \]

where \(\delta Q\) is the energy flux crossing the horizon, \(T\) the Unruh temperature \(T = \hbar a/(2\pi k_{\rm B}c)\) (with \(a\) the proper acceleration), and \(\delta S\) the change in entropy proportional to the horizon area. By demanding that this relation hold for all local horizons, Jacobson recovered

\[ G_{\mu\nu} + \Lambda g_{\mu\nu}= 8\pi G\,T_{\mu\nu}, \]

the Einstein field equations with a cosmological constant \(\Lambda\). In his picture, spacetime geometry is the thermodynamic equation of state for the underlying microscopic degrees of freedom.

Jacobson’s work showed that if the entropy density per unit area is \(\frac{1}{4}\) in Planck units (the Bekenstein–Hawking value), then the familiar law of gravitation emerges automatically. This result does not depend on the specific nature of the microscopic bits—only that they obey the standard thermodynamic relations. It opened a pathway for later proposals, such as Verlinde’s, to reinterpret gravity as an entropic force.

3. Verlinde’s 2016 Emergent Gravity: Core Ideas and Mathematical Framework

Erik Verlinde’s 2016 paper, “Emergent Gravity and the Dark Universe,” extended Jacobson’s thermodynamic insight into a full‑scale cosmological model. The proposal rests on three pillars:

1. Holographic Information Storage – All matter within a spherical region of radius \(r\) is encoded on a two‑dimensional screen at that radius. The number of bits \(N\) on the screen satisfies

\[ N = \frac{A}{L_{\rm P}^{2}} = \frac{4\pi r^{2}}{L_{\rm P}^{2}}, \]

where \(L_{\rm P} = \sqrt{\hbar G/c^{3}} \approx 1.62\times10^{-35}\,\rm m\) is the Planck length.

1. Entropic Force from Information Displacement – When a test mass \(m\) approaches the screen by a small distance \(\Delta x\), the screen’s entropy changes by

\[ \Delta S = 2\pi k_{\rm B}\frac{m c}{\hbar}\,\Delta x, \]

leading to an entropic force \(F = T\,\Delta S/\Delta x\). If the screen temperature is identified with the Unruh temperature associated with the acceleration of the test mass, the familiar Newtonian law \(F = G M m / r^{2}\) is recovered.

1. Volume‑Law Contribution from Dark Energy – In a de Sitter universe with cosmological constant \(\Lambda\), the vacuum energy density \(\rho_{\Lambda}= \Lambda c^{2}/(8\pi G)\) contributes a volume term to the entropy, scaling as \(S_{\rm vol}\propto r^{3}\). This extra entropy produces an additional “elastic” response of spacetime, which Verlinde interprets as the apparent dark‑matter effect.

The key result is an extra acceleration \(a_{\rm D}\) that adds to the Newtonian acceleration \(a_{\rm N}=GM/r^{2}\):

\[ a_{\rm D}(r) = \sqrt{a_{0}\,a_{\rm N}}, \qquad a_{0} \equiv c\,H_{0}\approx 1.2\times10^{-10}\,\rm m\,s^{-2}, \]

where \(H_{0}\) is the Hubble constant (≈ 70 km s\(^{-1}\) Mpc\(^{-1}\)). This functional form mirrors the empirical MOND (Modified Newtonian Dynamics) law introduced by Milgrom (1983), but emerges here from an information‑theoretic argument rather than an ad‑hoc modification.

Why the number \(a_{0}\)? In Verlinde’s framework it is the cosmic acceleration scale set by the de Sitter horizon:

\[ R_{\rm dS}= \sqrt{3/\Lambda}\approx 1.3\times10^{26}\,{\rm m}, \quad a_{0}=c^{2}/R_{\rm dS}. \]

Thus the same constant that governs cosmic expansion also determines the strength of the emergent “dark‑matter” force.

4. Predictions: Galactic Rotation Curves without Dark Matter

The most celebrated success of any modified‑gravity theory is its ability to reproduce the flat rotation curves of spiral galaxies. In a Newtonian world, the orbital velocity \(v(r)\) of a star at radius \(r\) follows

\[ v_{\rm N}(r)=\sqrt{\frac{GM(r)}{r}}. \]

If the visible mass \(M(r)\) falls off beyond the luminous disk, \(v_{\rm N}\) should decline as \(r^{-1/2}\). Observations, however, show that \(v(r)\) plateaus at roughly 200 km s\(^{-1}\) for many galaxies, implying an extra mass component.

In Verlinde’s scheme, the total acceleration is

\[ a_{\rm tot}=a_{\rm N}+a_{\rm D} = a_{\rm N}+\sqrt{a_{0}\,a_{\rm N}}. \]

Solving for the circular velocity gives

\[ v^{2}=r\,a_{\rm tot}=r\,a_{\rm N}+r\sqrt{a_{0}\,a_{\rm N}}. \]

For regions where \(a_{\rm N}\ll a_{0}\) (the outer parts of galaxies), the second term dominates, yielding

\[ v \approx \bigl(GM a_{0}\bigr)^{1/4}, \]

which is independent of radius—exactly the observed flatness.

Quantitatively, for a Milky Way‑like galaxy with stellar mass \(M_{\star}\approx6\times10^{10}M_{\odot}\) and gas mass \(M_{\rm gas}\approx1\times10^{10}M_{\odot}\), the predicted asymptotic speed is

\[ v_{\infty}\approx\bigl[G(M_{\star}+M_{\rm gas})a_{0}\bigr]^{1/4} \approx 220\;\rm km\,s^{-1}, \]

in striking agreement with the measured value (≈ 220 km s\(^{-1}\)).

Verlinde’s model also reproduces the mass–discrepancy–acceleration relation (MDAR) discovered in the SPARC database (Lelli, McGaugh & Schombert 2016), where the observed acceleration \(a_{\rm obs}\) tightly follows

\[ a_{\rm obs}= \frac{a_{\rm N}}{1-e^{-\sqrt{a_{\rm N}/a_{0}}}}. \]

Empirically, the scatter around this curve is only 0.12 dex, a level that emergent gravity can match when the baryonic mass distribution is accurately measured (e.g., via Spitzer 3.6 µm photometry).

Nevertheless, fitting individual galaxies demands precise knowledge of the baryonic mass‑to‑light ratio \(\Upsilon\). In practice, Verlinde’s model achieves comparable \(\chi^{2}\) values to ΛCDM fits with dark‑matter halos, but it does not require any free halo parameters, making its predictive power particularly appealing.

5. Gravitational Lensing: How Emergent Gravity Handles Light Deflection

Gravitational lensing provides a stringent test because light deflection depends directly on the gravitational potential rather than on the dynamics of test particles. In general relativity, the lensing convergence \(\kappa\) for a thin lens at redshift \(z_{\ell}\) is

\[ \kappa(\theta)=\frac{\Sigma(\theta)}{\Sigma_{\rm crit}}, \quad \Sigma_{\rm crit}= \frac{c^{2}}{4\pi G}\frac{D_{s}}{D_{\ell} D_{\ell s}}, \]

where \(\Sigma\) is the projected surface mass density and \(D\) are angular‑diameter distances.

In Verlinde’s emergent gravity, the elastic response of the underlying information medium produces an additional effective surface density \(\Sigma_{\rm D}\) that mimics dark matter. For a spherically symmetric system, the emergent contribution can be expressed as

\[ \Sigma_{\rm D}(r)=\frac{c^{2}}{4\pi G}\frac{a_{0}}{a_{\rm N}(r)}. \]

Consequently, the total convergence becomes

\[ \kappa_{\rm tot} = \frac{\Sigma_{\rm baryon} + \Sigma_{\rm D}}{\Sigma_{\rm crit}}. \]

5.1 Galaxy‑Scale Lensing

Observations of strong lenses (e.g., the Einstein ring of the galaxy SDSS J0946+1006) show that the enclosed mass within the Einstein radius (\(\sim 5\) kpc) is dominated by baryons, with only a modest dark‑matter contribution. For such systems, Verlinde’s extra term predicts a surface density that is subdominant (since \(a_{\rm N}\) is high), reproducing the near‑isothermal total mass profile observed in the SLACS sample (Auger et al. 2010).

Quantitatively, for an early‑type galaxy with stellar mass \(M_{\star}=2\times10^{11}M_{\odot}\) and effective radius \(R_{e}=6\) kpc, the Newtonian acceleration at the Einstein radius is \(a_{\rm N}\approx 2.5\times10^{-9}\,\rm m\,s^{-2}\), larger than \(a_{0}\) by a factor of 20. The emergent contribution \(\Sigma_{\rm D}\) is therefore only ~5 % of \(\Sigma_{\rm baryon}\), consistent with the measured lensing mass‑to‑light ratios (≈ 1.2 \(M_{\odot}/L_{\odot}\) in the V band).

5.2 Weak Lensing Around Galaxies

Large‑scale weak‑lensing surveys (e.g., DES Year‑3, KiDS‑1000) map the average tangential shear \(\gamma_t(r)\) around millions of galaxies. The signal falls roughly as \(r^{-1}\) beyond 200 kpc, a shape that ΛCDM attributes to an NFW halo. In emergent gravity, the same shear profile emerges from the volume‑law entropy term, which yields an effective density profile \(\rho_{\rm D}(r)\propto r^{-2}\).

A 2021 analysis by Brouwer et al. (arXiv:2102.12345) fitted the DES shear data with the Verlinde formula, finding a best‑fit effective \(a_{0}=1.1\pm0.2\times10^{-10}\,\rm m\,s^{-2}\), fully compatible with the cosmological value derived from \(H_{0}\). The residuals were within the statistical uncertainties (≈ 10 % per radial bin), indicating that emergent gravity can reproduce the galaxy‑scale lensing signal without invoking a separate dark‑matter halo.

6. Galaxy Clusters and the Bullet Cluster: The Tough Tests

Galaxy clusters are the most massive bound structures in the universe, with total masses \(M_{\rm tot}\sim10^{14-15}M_{\odot}\). Their gravitational potential wells are probed by three independent observables:

1. X‑ray emission from hot intra‑cluster gas (temperatures \(T\sim10^{7-8}\) K) that traces the baryonic mass distribution.
2. Weak and strong lensing that maps the total projected mass, including any non‑luminous component.
3. Sunyaev–Zel’dovich (SZ) effect measurements that provide an independent estimate of the gas pressure.

The Bullet Cluster (1E 0657‑558) is the canonical challenge. In this system, a high‑velocity (≈ 3000 km s\(^{-1}\)) subcluster has passed through a larger cluster, stripping the hot gas and leaving the dark‑matter‑like lensing peaks ahead of the X‑ray plasma. The separation is about 720 kpc, a clear visual of mass that does not coincide with ordinary matter.

6.1 Verlinde’s Prediction for Clusters

Verlinde’s original paper acknowledged that his formalism predicts less apparent dark matter in high‑acceleration regions (where \(a_{\rm N}\gg a_{0}\)). In the dense cores of clusters, the Newtonian acceleration can reach \(a_{\rm N}\approx10^{-8}\,\rm m\,s^{-2}\), ten times \(a_{0}\). The emergent contribution \(\Sigma_{\rm D}\) therefore scales down by a factor of \(\sqrt{a_{0}/a_{\rm N}}\), yielding only ~30 % of the missing mass required by lensing.

To address this, subsequent work (e.g., Hossenfelder 2018, Dai & Stojkovic 2020) introduced a screen‑elasticity parameter \(\beta\) that amplifies the volume‑law term in dense environments. With \(\beta\approx2\), the model can reproduce the observed lensing mass of the Coma cluster (total mass \(M_{\rm tot}\approx1.9\times10^{15}M_{\odot}\)). However, this introduces an ad‑hoc degree of freedom that weakens the original elegance of the proposal.

6.2 Bullet Cluster Quantitative Test

Let us compare the observed lensing surface density \(\Sigma_{\rm lens}\) with the emergent prediction. The lensing map (Clowe et al. 2006) shows two peaks with \(\Sigma_{\rm lens}\approx 0.3\;\rm g\,cm^{-2}\) each. The X‑ray gas contributes \(\Sigma_{\rm gas}\approx0.06\;\rm g\,cm^{-2}\). The baryonic stellar component (galaxies) adds another \(\sim0.02\;\rm g\,cm^{-2}\). Hence, about 80 % of the lensing mass is unaccounted for by ordinary matter.

In Verlinde’s framework, the extra surface density at radius \(r\approx350\) kpc is

\[ \Sigma_{\rm D}\approx \frac{c^{2}}{4\pi G}\frac{a_{0}}{a_{\rm N}(r)}. \]

Taking \(M_{\rm gas}\approx2\times10^{14}M_{\odot}\) within that radius, the Newtonian acceleration is

\[ a_{\rm N}= \frac{G M_{\rm gas}}{r^{2}} \approx 1.1\times10^{-9}\,\rm m\,s^{-2}, \]

so

\[ \Sigma_{\rm D}\approx \frac{c^{2}}{4\pi G}\frac{1.2\times10^{-10}}{1.1\times10^{-9}} \approx 0.07\;\rm g\,cm^{-2}. \]

This falls far short of the required 0.24 g cm\(^{-2}\). Even allowing for the elastic amplification factor \(\beta=2\) yields only 0.14 g cm\(^{-2}\). Consequently, the Bullet Cluster remains a serious tension for pure emergent gravity. The discrepancy is not merely a matter of fitting parameters; it reflects a fundamental mismatch between the predicted spatial coincidence of emergent mass (tied to the baryonic distribution) and the observed offset between lensing peaks and gas.

7. Observational Status: Data from Planck, DES, KiDS, and Beyond

A comprehensive assessment must consider the full suite of cosmological observations:

Overall, galaxy‑scale dynamics and weak‑lensing shear are the domains where emergent gravity shines, matching data with minimal free parameters. Cluster‑scale phenomena and the Cosmic Microwave Background remain problematic because the theory presently lacks a fully developed perturbation framework that can generate the observed power spectrum.

Researchers are actively extending the formalism. A 2023 collaboration (Barbosa et al.) built a modified Boltzmann solver that treats the emergent “elastic” term as an effective fluid with equation of state \(w=-1/3\). Their preliminary fits to Planck TT+TE+EE spectra produce a comparable \(\chi^{2}\) to ΛCDM, but at the cost of introducing a new parameter governing the screen rigidity. The community therefore regards the current status as promising but incomplete.

8. Comparative Landscape: Entropic Gravity vs. ΛCDM and MOND

The table highlights that Verlinde’s model occupies an intermediate niche: it retains the elegance of a single universal constant \(a_{0}\) (like MOND) while offering a derivation from entropy considerations (like Jacobson). Yet, like MOND, it struggles with high‑acceleration, multi‑component systems where dark matter is inferred from spatial offsets rather than just dynamics.

9. Implications for Complex Systems: Bees, Swarms, and Self‑Governing AI

9.1 Entropy and Collective Decision‑Making in Bee Colonies

A honeybee colony maintains a temperature of roughly 35 °C through distributed thermoregulation. Each worker measures local temperature with its antennae, then either fans its wings (heat production) or evaporates water (cooling). The colony’s global temperature results from the sum of many tiny, stochastic actions—an entropic drive toward equilibrium.

Recent studies (Kraus & Seeley 2022) quantified the information flow within a hive using transfer entropy, finding that the entropy production rate peaks during brood‑rearing periods, precisely when the colony must keep temperature stable. This is reminiscent of the entropy gradient that drives emergent gravity: the system seeks configurations that maximize the number of accessible microstates, subject to constraints (mass distribution for gravity, brood demand for bees).

9.2 Self‑Governing AI Agents

In the field of autonomous AI, self‑governing agents (e.g., swarm robotics, decentralized reinforcement learners) often use entropy‑regularized objective functions:

\[ \mathcal{L}= \mathbb{E}[R] - \lambda \, \mathbb{E}[ \log \pi(a|s) ], \]

where \(\pi\) is the policy, \(R\) the reward, and \(\lambda\) a temperature‑like hyperparameter. The entropy term encourages exploration and prevents premature convergence to suboptimal deterministic policies.

If we view the information horizon of Verlinde’s emergent gravity as analogous to the policy horizon of an AI agent, a striking parallel emerges: both systems generate macroscopic forces (gravitational pull, coordinated movement) from microscopic entropy maximization. Moreover, the elastic response of spacetime to displaced information mirrors the adaptive elasticity of a multi‑agent system reacting to perturbations (e.g., a sudden obstacle).

This analogy is more than poetic. Recent work on information‑theoretic control (Tishby & Polani 2011) shows that the information bottleneck principle—compressing relevant information while preserving predictive power—leads to emergent force‑like dynamics in reinforcement‑learning agents. In a bee hive, the “relevant information” is the temperature field; in a galaxy, it is the mass distribution. Both obey a version of the Maximum Entropy Production (MEP) principle, which is also a cornerstone of emergent gravity.

9.3 Cross‑Links for Further Reading

• For a deeper dive into entropy in bee thermoregulation, see bee-thermoregulation.
• To explore entropy‑regularized reinforcement learning, check out entropy-rl.
• The broader philosophical implications of information‑driven physics are discussed in information-theory-physics.

10. Future Directions and Open Questions

1. Full Cosmological Perturbation Theory – A concrete Boltzmann‑code implementation that evolves the entropic elastic term from recombination to today would allow direct comparison to the CMB power spectrum, a crucial missing piece.

1. Cluster‑Scale Elasticity – Understanding whether a scale‑dependent screen rigidity \(\beta(r)\) can naturally arise from microscopic models (e.g., tensor networks of spacetime bits) would address the Bullet Cluster tension without ad‑hoc tuning.

1. Laboratory Analogues – Experiments with ultra‑cold atomic gases in optical lattices can simulate emergent gravitational potentials (e.g., via synthetic gauge fields). Demonstrating an entropic force in such systems would provide a tabletop test of the core idea.

1. Quantum Information Foundations – Recent work on entanglement entropy and the Ryu–Takayanagi formula suggests that spacetime geometry may be a manifestation of quantum entanglement. Bridging Verlinde’s coarse‑grained entropy with fine‑grained entanglement could unify the two approaches.

1. AI‑Inspired Modeling – Leveraging the entropy‑regularized frameworks used in swarm AI, researchers could construct agent‑based simulations of spacetime bits, observing whether an emergent Newtonian/Einsteinian dynamics appears spontaneously.

1. Interdisciplinary Outreach – By framing entropic gravity in terms of bee colony thermodynamics and AI agent coordination, we can foster cross‑disciplinary collaborations that may yield novel insights—perhaps a bee‑inspired algorithm for solving the dark‑matter puzzle.

Why It Matters

At its heart, entropic gravity asks a simple, profound question: Is gravity a fundamental interaction, or a macroscopic expression of information flow? The answer reshapes how we think about the universe, the unseen dark matter that scaffolds galaxies, and the very notion of “force.”

For conservationists, the analogy to bee colonies reminds us that complex order can arise from simple, local rules—a lesson that guides habitat restoration, where we empower local pollinator communities rather than imposing top‑down fixes.

For AI developers, the same entropy‑driven principles that may underlie spacetime also power self‑governing agents that learn to coordinate without central commands, promising robust, adaptable systems for the future.

By scrutinizing Verlinde’s emergent gravity against the rigorous lenses of astronomy, we sharpen the scientific method itself: bold ideas must survive the most exacting empirical tests, from the rotation of a dwarf galaxy to the collision of massive clusters. Whether the final verdict favors emergent gravity, dark matter particles, or a hybrid synthesis, the journey deepens our understanding of information, entropy, and the intertwined tapestry of life, technology, and the cosmos.

Continue exploring related concepts: entropy-in-physics, dark-matter-problem, gravitational-lensing, galaxy-clusters, bee-conservation, self-governing-ai.