← Back to Part VIII: Lagrangian Connections

Synthesis: The Master Lagrangian

Six Lagrangian structures unified into a single variational framework for gravitational physics

1. The Master Lagrangian

The complete variational principle for asymptotically flat general relativity, incorporating all scales from the spatial bulk to the celestial sphere, is the six-term Master Lagrangian:

$$S_{\rm master} = \underbrace{\frac{1}{16\pi G}\int_{\mathcal{M}} R\,\sqrt{-g}\,d^4x}_{\text{I. Einstein--Hilbert (bulk)}}$$

$$+\; \underbrace{\frac{1}{8\pi G}\oint_{\partial\mathcal{M}} K\,\sqrt{h}\,d^3y}_{\text{II. GHY (boundary)}} \;+\; \underbrace{\int_\Sigma \bigl(R_\Sigma + |\nabla f|^2\bigr) e^{-f}\,d\mu_g}_{\text{III. Perelman } \mathcal{F} \text{ (spatial slice)}}$$

$$+\; \underbrace{\frac{k}{4\pi}\int_{\mathcal{H}} \mathrm{tr}\!\left(A \wedge dA + \tfrac{2}{3} A \wedge A \wedge A\right)}_{\text{IV. Chern--Simons (near-horizon)}}$$

$$+\; \underbrace{\int_{\mathscr{I}^+} \bigl(N_{AB} N^{AB} + \partial_u m_B\bigr)\, du\, d^2\Omega}_{\text{V. Bondi--Sachs } (\mathscr{I}^+)} \;+\; \underbrace{\mathcal{K}(\omega_{S^2})}_{\text{VI. Mabuchi } (S^2)}$$

2. Role of Each Term

Einstein–Hilbert (bulk): The fundamental action for spacetime geometry. Its Euler–Lagrange equations give$G_{\mu\nu} = 8\pi G\, T_{\mu\nu}$. Variation produces the Palatini identity$\delta R_{\mu\nu} = \nabla_\rho \delta\Gamma^\rho{}_{\mu\nu} - \nabla_\nu \delta\Gamma^\rho{}_{\mu\rho}$.
GHY boundary term: Ensures a well-posed Dirichlet variational problem. The extrinsic curvature $K$ cancels the normal derivative boundary terms from $\delta R$. Under 3+1 decomposition, it becomes the ADM canonical momentum $\pi^{ij} = \sqrt{h}(K h^{ij} - K^{ij})$.
Perelman $\mathcal{F}$ (spatial slice): The Ricci flow action principle on the spatial slice $\Sigma$. Its gradient flow drives geometrization. The dilaton $f$ plays the role of the conformal factor, and critical points are gradient Ricci solitons.
Chern–Simons (near-horizon): Captures the topological sector near black hole horizons. The connection $A$ is the$\mathrm{SO}(2,1)$ or $\mathrm{SL}(2,\mathbb{R})$ gauge field from the near-horizon geometry. Wilson lines encode the reduced volume, and the CS level $k = A_H / (4G)$ gives the Bekenstein–Hawking entropy.
Bondi–Sachs ($\mathscr{I}^+$): The action at null infinity whose symplectic structure encodes gravitational radiation. The news tensor $N_{AB}$ is the radiative degree of freedom, and the Bondi mass loss$dm_B/du \leq 0$ follows from the equations of motion.
Mabuchi K-energy ($S^2$): The Kähler-geometric functional on the celestial sphere. Its gradient flow (KRF) drives the conformal metric toward round. Spin memory appears as a K-energy displacement$\Delta\mathcal{K} \neq 0$.

3. The Master Equation

The six terms are not independent. Their gradient flows are linked by the Master Equation chain, which relates the monotonicity of each functional to the others:

$$\frac{d\mathcal{F}}{dt} = -\frac{d\mathcal{K}}{dt}\bigg|_{S^2} \;\geq\; |\Delta\Psi|^2 \cdot \frac{c_{\rm eff}}{12\pi}$$

$$= -\frac{dm_B}{du} \cdot 8G = \frac{d\widetilde{V}^{-1}}{d\tau} \cdot (4\pi\tau)^{n/2}\, e^{-\ell_{\min}}$$

Reading left to right, this chain states:

The increase of Perelman's $\mathcal{F}$ on the spatial slice equals the decrease of the Mabuchi K-energy on $S^2$: geometrization of $\Sigma$ and uniformization of$S^2$ are dual processes.
The K-energy decrease is bounded below by the spin memory squared, weighted by the effective central charge $c_{\rm eff}$ of the boundary CFT.
This equals $8G$ times the Bondi mass loss rate: energy radiated to infinity drives both the bulk geometrization and the celestial uniformization.
The far right relates the Bondi mass loss to the reduced volume $\widetilde{V}$ and the $\mathcal{L}$-length minimizer $\ell_{\min}$, connecting to Perelman's monotonicity formulas.

4. Variational Derivation

The Master Equation follows from the first variation of $S_{\rm master}$. Consider a one-parameter family of asymptotically flat metrics $g(\epsilon)$. The total derivative is:

$$\frac{dS_{\rm master}}{d\epsilon}\bigg|_{\epsilon=0} = \int_{\mathcal{M}} G_{\mu\nu}\, \delta g^{\mu\nu}\, d^4x + \text{(boundary terms at } \partial\mathcal{M}, \mathscr{I}^+, S^2\text{)}$$

On shell ($G_{\mu\nu} = 0$ in vacuum), only boundary terms survive. The GHY term cancels the timelike boundary contribution, leaving:

$$0 = \delta\mathcal{F}\big|_\Sigma + \delta S_{\rm CS}\big|_{\mathcal{H}} + \delta S_{\rm BS}\big|_{\mathscr{I}^+} + \delta\mathcal{K}\big|_{S^2}$$

Taking the time derivative and using the monotonicity of each functional along its gradient flow yields the Master Equation chain. The key identity linking the Perelman and Bondi sectors is:

$$2\int_\Sigma |R_{ij} + \nabla_i\nabla_j f|^2\, e^{-f}\, d\mu = \frac{1}{8\pi G}\oint_{S^2_\infty} N_{AB} N^{AB}\, d\Omega$$

which identifies the Perelman monotonicity integrand with the Bondi news flux at spatial infinity.

5. BV Cohomology of the Master Action

The BV formalism applied to $S_{\rm master}$ yields a master action satisfying:

$$(S_{\rm BV}^{\rm master},\; S_{\rm BV}^{\rm master}) = 0$$

The cohomology $H^*(S_{\rm BV}^{\rm master})$ decomposes by the region of support:

$$H^0(s)\big|_\Sigma = \{\mathcal{F}, \mathcal{W}, \widetilde{V}\}, \qquad H^0(s)\big|_{\mathscr{I}^+} = \{m_B, J_B, Q_f, Q_Y\}$$

$$H^0(s)\big|_{S^2} = \{\mathcal{K}, \text{Futaki}\}, \qquad H^1(s) = \{\Delta\Psi, \Delta C_{AB}\}$$

The $H^0$ classes are gauge-invariant observables in each sector. The $H^1$ classes are the memory effects: displacement memory $\Delta C_{AB}$ and spin memory $\Delta\Psi$. The isomorphism between sectors is the content of the infrared triangle.

6. The Open Synthesis Problem

Central Question

Does there exist a single functional $\mathcal{S}[g]$ on the space of asymptotically flat 4-metrics such that:

Its restriction to a spatial slice $\Sigma$ equals the Perelman functional:$\;\mathcal{S}\big|_\Sigma = \mathcal{F}(g_\Sigma, f)$
Its restriction to null infinity equals the Bondi–Sachs action:$\;\mathcal{S}\big|_{\mathscr{I}^+} = S_{\rm BS}$
Its restriction to the celestial sphere equals the Mabuchi K-energy:$\;\mathcal{S}\big|_{S^2} = \mathcal{K}(\omega_{S^2})$
Its gradient flow simultaneously drives geometrization of $\Sigma$, zero news at $\mathscr{I}^+$, and round celestial sphere

Formally, such a functional would satisfy:

$$\frac{\delta \mathcal{S}}{\delta g_{\mu\nu}} = 0 \quad \Longleftrightarrow \quad \begin{cases} R_{ij} + \nabla_i\nabla_j f = 0 & \text{on } \Sigma \\ N_{AB} = 0 & \text{on } \mathscr{I}^+ \\ R_{z\bar{z}} = g_{z\bar{z}} & \text{on } S^2 \end{cases}$$

The Master Lagrangian $S_{\rm master}$ presented above is a candidate, but it is a sum of six terms with independent variational principles. The open problem asks whether a single, irreducible functional exists. Progress would require:

A unified geometric structure interpolating between Riemannian, Lorentzian, and Kähler geometry
A holographic renormalization scheme relating bulk and boundary terms
A proof that the Master Equation chain is not merely a consequence of the Einstein equations but an independent variational identity

$$\mathcal{S}[g] \;\stackrel{?}{=}\; \frac{1}{16\pi G}\int_{\mathcal{M}} \bigl(R + \Lambda_{\rm eff}[g]\bigr)\sqrt{-g}\, d^4x$$

where $\Lambda_{\rm eff}[g]$ is a nonlocal functional encoding the Perelman, Bondi–Sachs, and Mabuchi sectors

7. Consistency Checks

The Master Lagrangian must satisfy several nontrivial consistency conditions. We verify three:

7.1 Dimensional Reduction

Under the ADM decomposition $g_{\mu\nu} \to (N, N^i, h_{ij})$, the Einstein–Hilbert plus GHY terms reduce to:

$$S_{\rm EH} + S_{\rm GHY} = \int dt \int_\Sigma \bigl(\pi^{ij}\dot{h}_{ij} - N\mathcal{H} - N^i\mathcal{H}_i\bigr)\, d^3x$$

The Hamiltonian constraint $\mathcal{H} = 0$ and momentum constraint $\mathcal{H}_i = 0$must be compatible with the Perelman sector. Under Wick rotation $N \to iN_E$, the Hamiltonian constraint becomes $R_\Sigma - |K|^2 + K^2 = 0$, which at the level of the Perelman functional gives:

$$\mathcal{H}_{\rm Perelman} = R_\Sigma + |\nabla f|^2 - \frac{n}{2\tau} = 0 \quad \text{(dilaton constraint)}$$

7.2 Asymptotic Matching

At large $r$, the Bondi–Sachs term must match the asymptotic expansion of the EH action. In Bondi coordinates $(u, r, \theta^A)$:

$$\lim_{r \to \infty} r^2 \int R\,\sqrt{-g}\, du\, d^2\Omega = \int_{\mathscr{I}^+} \bigl(N_{AB} N^{AB} + 2\partial_u m_B\bigr)\, du\, d^2\Omega$$

7.3 BV Closure

The classical master equation $(S_{\rm BV}^{\rm master}, S_{\rm BV}^{\rm master}) = 0$ requires that the BRST transformations in the DeTurck and BMS sectors are compatible. This holds because both derive from the same underlying diffeomorphism invariance of the Einstein equations, restricted to different regions of spacetime.

8. Symmetry Reductions

The Master Lagrangian admits natural reductions under symmetry assumptions. For spherically symmetric spacetimes, only the EH, GHY, and Perelman terms contribute nontrivially:

$$S_{\rm master}^{\rm sph} = \frac{1}{4G}\int \bigl(1 - 2m'(r)\bigr)\, dr + \int_\Sigma \bigl(R_\Sigma + |\nabla f|^2\bigr) e^{-f}\, d\mu$$

The Bondi–Sachs and Mabuchi terms vanish because spherical symmetry implies zero news ($N_{AB} = 0$) and a round celestial sphere ($\mathcal{K} = 0$). The CS term contributes only if there is a horizon, giving the entropy $S = A_H/(4G)$.

For axisymmetric radiating spacetimes, all six terms are active. The Bondi mass loss becomes:

$$\frac{dm_B}{du} = -\frac{1}{8\pi G}\sum_{\ell \geq 2} \frac{(\ell+2)!}{(\ell-2)!}\, |N_\ell(u)|^2$$

where $N_\ell$ are the multipole moments of the news. The $\ell = 2$ quadrupole dominates for compact binary mergers.

9. Outlook

The Master Lagrangian framework reveals that general relativity possesses far richer variational structure than the Einstein–Hilbert action alone suggests. The six terms are not merely formal additions but reflect genuine physical phenomena:

The Perelman sector governs the long-time behavior of spatial geometry (geometrization)
The Chern–Simons sector captures the topological features of horizons (entropy)
The Bondi–Sachs sector encodes gravitational radiation (energy loss)
The Mabuchi sector controls the asymptotic conformal geometry (uniformization)

The Master Equation chain binds these into a single dynamical narrative: as a spacetime evolves, energy flows from the bulk to null infinity, driving simultaneous geometrization of spatial slices and uniformization of the celestial sphere. Gravitational memory effects are the permanent records of this process, living in the first BV cohomology group.

$$\text{Geometrization} + \text{Radiation} + \text{Uniformization} = \text{One Gradient Flow}$$

Simulation: The Master Equation Chain

Python

script.py113 lines

import numpy as np
import matplotlib.pyplot as plt

# The Master Equation Chain
# Simulate ALL five monotone quantities simultaneously:
# 1. Perelman W-entropy (increasing)
# 2. Zamolodchikov c-function (decreasing)
# 3. Bondi mass (decreasing)
# 4. Spin memory (accumulating)
# 5. Reduced volume (decreasing)

t = np.linspace(0, 10, 1000)
dt = t[1] - t[0]

# Model a "merger event" driving all flows
t_event = 3.0
sigma_event = 1.0
event_rate = np.exp(-(t - t_event)**2 / (2 * sigma_event**2))
event_rate += 0.3 * np.exp(-(t - 7)**2 / (2 * 0.5**2))

# 1. Perelman W-entropy: dW/dt = 2 * int |Ric + Hess(f)|^2 >= 0
dW_dt = 2.0 * event_rate**2 + 0.02
W_entropy = -5.0 + np.cumsum(dW_dt) * dt

# 2. Zamolodchikov c-function: dc/dt <= 0
c_0 = 12.0
dc_dt = -1.5 * event_rate**2 - 0.01
c_function = c_0 + np.cumsum(dc_dt) * dt

# 3. Bondi mass: dm_B/du = -(1/8) N_AB N^AB <= 0
M_0 = 1.0
dm_dt = -0.8 * event_rate**2
m_bondi = M_0 + np.cumsum(dm_dt) * dt

# 4. Spin memory: accumulates monotonically
d_spin_dt = 0.5 * event_rate * np.sin(4 * t)
spin_memory = np.cumsum(np.abs(d_spin_dt)) * dt

# 5. Reduced volume: dV_tilde/dtau <= 0
V_0 = (4 * np.pi)**1.5
dV_dt = -0.3 * event_rate**2 - 0.005 * np.exp(-0.1 * t)
V_reduced = V_0 + np.cumsum(dV_dt) * dt

fig, axes = plt.subplots(5, 1, figsize=(12, 14), sharex=True)
fig.patch.set_facecolor('#0a0a1a')
fig.suptitle('The Master Equation Chain: Five Parallel Monotonicities',
             color='#e2e8f0', fontsize=16, fontweight='bold', y=0.98)

colors = ['#38bdf8', '#a78bfa', '#fb923c', '#e879f9', '#34d399']
titles = [
    'Perelman W-Entropy (INCREASING)',
    'Zamolodchikov c-Function (DECREASING)',
    'Bondi Mass m_B (DECREASING)',
    'Spin Memory |Delta Psi| (ACCUMULATING)',
    'Reduced Volume V-tilde (DECREASING)'
]
data = [W_entropy, c_function, m_bondi, spin_memory, V_reduced]
ylabels = ['W', 'c(t)', 'm_B', '|Delta Psi|', 'V-tilde']
arrows = ['UP', 'DOWN', 'DOWN', 'UP', 'DOWN']

for i, ax in enumerate(axes):
    ax.set_facecolor('#0a0a1a')
    ax.tick_params(colors='#94a3b8')
    for spine in ax.spines.values():
        spine.set_color('#334155')
    ax.grid(True, alpha=0.1, color='#334155')

ax.plot(t, data[i], color=colors[i], linewidth=2.5)
    ax.fill_between(t, data[i], data[i].min() if arrows[i] == 'UP' else data[i].max(),
                     alpha=0.08, color=colors[i])
    ax.set_ylabel(ylabels[i], color=colors[i], fontsize=12, fontweight='bold')

ax.text(0.98, 0.85, titles[i], transform=ax.transAxes, color=colors[i],
            fontsize=11, fontweight='bold', ha='right', va='top',
            bbox=dict(boxstyle='round,pad=0.3', facecolor='#0a0a1a', edgecolor=colors[i], alpha=0.8))

ax.axvline(x=t_event, color='#f59e0b', linestyle=':', alpha=0.3)
    ax.axvline(x=7.0, color='#f59e0b', linestyle=':', alpha=0.2)

mid_idx = len(t) // 2
    y_mid = data[i][mid_idx]
    arrow_dir = 0.08 * (data[i].max() - data[i].min()) * (1 if arrows[i] == 'UP' else -1)
    ax.annotate('', xy=(t[mid_idx], y_mid + arrow_dir), xytext=(t[mid_idx], y_mid - arrow_dir),
                arrowprops=dict(arrowstyle='->', color=colors[i], lw=2.5))

axes[-1].set_xlabel('Flow time t', color='#94a3b8', fontsize=13)

axes[0].text(t_event, axes[0].get_ylim()[1] * 0.95, 'Primary event',
             color='#f59e0b', fontsize=9, ha='center', va='top')
axes[0].text(7.0, axes[0].get_ylim()[1] * 0.95, 'Secondary event',
             color='#f59e0b', fontsize=9, ha='center', va='top')

plt.tight_layout(rect=[0, 0, 1, 0.96])
plt.savefig('output.png', dpi=130, bbox_inches='tight', facecolor='#0a0a1a')

print("The Master Equation Chain: Five Parallel Monotonicities")
print("=" * 55)
print()
print("Quantity                  Initial     Final       Direction")
print("-" * 60)
print(f"Perelman W-entropy:      {W_entropy[0]:>8.3f}    {W_entropy[-1]:>8.3f}    INCREASING")
print(f"Zamolodchikov c:         {c_function[0]:>8.3f}    {c_function[-1]:>8.3f}    DECREASING")
print(f"Bondi mass m_B:          {m_bondi[0]:>8.3f}    {m_bondi[-1]:>8.3f}    DECREASING")
print(f"Spin memory |Psi|:       {spin_memory[0]:>8.3f}    {spin_memory[-1]:>8.3f}    ACCUMULATING")
print(f"Reduced volume V~:       {V_reduced[0]:>8.3f}    {V_reduced[-1]:>8.3f}    DECREASING")
print()
print("Master Equation Chain:")
print("  dW/dt = -dK/dt >= |Delta Psi|^2 * c/(12pi) = -dm_B/du * 8G = dV~^{-1}/dtau * ...")
print()
print("All five quantities exhibit the correct monotonicity,")
print("driven by the same underlying geometric event (GW burst / Ricci flow singularity).")
print("This demonstrates the parallel structure of the Master Lagrangian.")

Click Run to execute the Python code

Code will be executed with Python 3 on the server

Video Lecture

← Mabuchi & Kähler-Ricci Flow Back to Part VIII →