Does energy-parity subtraction render induced gravity finite?

Technical Summary: Does Energy-Parity Subtraction Render Induced Gravity Finite?

Problem Statement
The paper addresses a specific, fifty-year-old obstruction to Sakharov's induced gravity program. While Sakharov proposed that the Einstein–Hilbert action arises from vacuum fluctuations of matter fields, the program has stalled due to the "counterterm problem." When a physical regulator (a medium with a preferred rest frame) is used to cut off ultraviolet divergences, the induced effective action contains not only the desired covariant curvature term ($\int \sqrt{g} R$) but also large, non-covariant counterterms of order $\Lambda^4$ and $\Lambda^2$. These terms, dependent on the regulator's preferred frame (e.g., extrinsic curvature $K_{ij}$, acceleration $a_i$), swamp the Einstein–Hilbert term, rendering the induced gravity framework non-viable. The central question is whether a specific structural hypothesis about the vacuum can remove these power divergences without destroying the induced Newton constant.

Methodology
The authors propose a "soft $Z_2$ energy-parity doubling" of the vacuum. The quantum vacuum is modeled as a twin pair of scalar fields, $\Phi_+$ and $\Phi_-$, with masses $m^2_\pm = m^2 \mp \gamma$ (where $0 < \gamma \ll \Lambda^2$).

• Functional Integral: The fields enter the functional integral with opposite powers of the determinant (a subtraction scheme).
• Fakeon Prescription: The heavier partner ($\Phi_-$) is removed from the asymptotic spectrum via the "fakeon" prescription (Anselmi), ensuring it is not a ghost and preserving perturbative unitarity.
• Medium: The calculation is performed in a Euclidean signature with a physical medium possessing a rest frame and a $z=2$ dispersive elasticity ($\omega^2 = c^2k^2 + k^4/\Lambda^2$) to provide a physical UV cutoff.
• Core Mechanism: The effective action is computed as the difference of two one-loop traces: $\Gamma = \frac{1}{2}\text{Tr}\ln D_+ - \frac{1}{2}\text{Tr}\ln D_-$. This difference generates a "twin factor" in the proper-time integral: $e^{-sm^2_+} - e^{-sm^2_-} = 2e^{-sm^2}\sinh(s\gamma)$.

Key Contributions and Results

1. The Degree Lemma (Lemma 2):
   The central technical result is a lemma stating that the twin factor $2e^{-sm^2}\sinh(s\gamma) \approx 2s\gamma e^{-sm^2}$ supplies an extra power of the proper time $s$. In the heat-kernel expansion $\text{Tr} e^{-sD_0} \sim \sum a_k s^{k-2}$, this extra factor of $s$ lowers the degree of divergence of every sector by exactly two.

   • $a_0$ sector (Cosmological term): Divergence drops from $\Lambda^4$ to $\gamma \Lambda^2$.
   • $a_1$ sector (Newton's constant): Divergence drops from $\Lambda^2$ to $\ln(\Lambda^2/m^2)$.
   • $a_2$ sector (Curvature squared): Becomes finite (no cutoff dependence).
     Crucially, this reduction applies indiscriminately to both covariant terms (like $R$) and non-covariant terms (like $K_{ij}K^{ij}$), thereby removing the non-covariant counterterms that previously obstructed the theory.

2. Newton's Constant as a Residue:
   The induced Newton constant $G_N$ is identified as the residue of the $a_1$ term. The result is:
   $$G_N^{-1} = \frac{\gamma}{\pi} \left(\frac{1}{6} - \xi\right) \ln\left(\frac{\Lambda^2}{m^2}\right)$$
   where $\xi$ is the non-minimal coupling. The paper demonstrates that $G_N$ is finite and vanishes if the vacuum is perfectly balanced ($\gamma = 0$). Gravity is thus interpreted as the residue of the failure of the twin sectors to cancel.

3. The Entropy–Newton Identity:
   The authors compute the entanglement entropy ($S$) of the twin-subtracted vacuum across an arbitrary surface using the conical (replica) method. They find that the entropy is governed by the exact same integral $Q$ and the same coefficient $(\frac{1}{6} - \xi)$ as the Newton constant.
   $$S_{twin} = \frac{A}{4G_N}$$
   This is established as an identity between two finite quantities at one-loop and logarithmic order, rather than a relation between two divergent quantities. The factor $1/4$ emerges naturally from the coincidence of the conical excess ($1/6$) and the bulk coefficient $a_1$.

4. Sign Chain and Stability:
   A "sign chain" is established linking four physical conditions: $S > 0 \iff G_N > 0 \iff \xi < 1/6 \iff$ the Kaplan–Sundrum vacuum is stable. This implies that if the non-minimal coupling $\xi$ exceeds $1/6$, the theory simultaneously yields negative entropy, repulsive gravity, and an unstable vacuum.

5. Numerical Verification:
   The paper provides four independent numerical checks (reproducibility package):

   • Exact verification of the orbifold sum identity used in the conical calculation.
   • Confirmation that the logarithmic slope of the integral $Q$ matches the predicted $2\gamma$.
   • Demonstration that without the twin subtraction, a power divergence ($s^{-1/4}_{min}$) persists even with $z=2$ dispersion, proving finiteness is due to the twin, not the medium.
   • Sector-by-sector confirmation of the degree lemma.

Claims and Limitations

• What is Claimed: The paper claims to dissolve the specific obstruction of non-covariant counterterms in induced gravity. It argues that the perturbative non-renormalisability of Einstein–Hilbert as a fundamental action is a category error; in this framework, $R$ is an induced residue, and the question of its renormalisability is replaced by the finiteness of the twin-subtracted vacuum trace. It establishes that black-hole entropy and Newton's constant are two readings of the same finite integral.
• What is Not Claimed:
  • The authors explicitly state they do not claim a full quantum theory of gravity.
  • They do not solve the cosmological constant problem entirely; a residual term $\gamma \Lambda^2$ remains in the $a_0$ sector (the "ground-state problem" is delegated).
  • The results are strictly at one-loop and logarithmic order.
  • The explicit finite coefficients ($O(\Lambda^0)$) for the anisotropic cone are not computed. The authors provide a structural argument (Proposition 7) that the finite part is covariantly organized at the "calibration point" due to the entropy–action correspondence, but they do not calculate the specific numbers.
  • Whether the renormalization group flow actually reaches this calibration point is an open question addressed in companion work.

Conclusion
The paper posits that the "counterterm problem" of induced gravity is a technical issue solvable by a soft energy-parity doubling of the vacuum. By lowering the divergence degree of the heat-kernel sectors by two, the twin subtraction eliminates the non-covariant power divergences that previously swamped the Einstein–Hilbert term. Consequently, Newton's constant and horizon entropy emerge as finite, identical integrals, suggesting that the weakness of gravity is a consequence of the near-perfect cancellation of twin vacuum sectors rather than a fundamental scale hierarchy.