Asymptotically Anti-de Sitter Spacetimes

Negative cosmological constant makes conformal infinity timelike. That single causal fact reorganizes the initial-value problem: bulk data alone are insufficient, boundary conditions determine flux and evolution, asymptotic expansions separate prescribed boundary geometry from state data, and conserved charges depend on the same boundary phase space. The long-time consequences—reflection, resonance, trapping, and possible instability—remain model dependent.

Helpful background. Lorentzian Metrics and Causal Cones supplies boundary-normal classification; Geodesics and Jacobi Fields supplies null travel and trapping; Causality and Global Structure gives the volume route; Exact Solutions of Einstein’s Equations supplies AdS coordinates; Formulations of General Relativity supplies evolution; Gravitational Actions and Boundary Terms supplies variational data.

Timelike infinity changes predictability

Global AdS$_{d+1}$ has metric

$$d\widetilde s^2 =\frac{\ell^2}{\cos^2\rho} \left(-dt^2+d\rho^2+\sin^2\rho\,d\Omega_{d-1}^2\right), \qquad \Lambda=-\frac{d(d-1)}{2\ell^2}.$$

Multiplication by $\Omega^2=(\cos\rho/\ell)^2$ extends the unphysical metric to $\rho=\pi/2$. The normal is spacelike, so the boundary is timelike and has conformal topology $\mathbb R\times S^{d-1}$ on the universal cover. Radial light crosses from the center to the boundary in global time $\pi/2$, although its physical affine parameter diverges.

Timelike Conformal Infinity in AdS develops this geometry, distinguishes the universal cover from the closed-time hyperboloid, and separates global from Poincaré-patch boundaries. Why AdS Is Not Globally Hyperbolic then gives the compact-diamond proof: sequences in a causal diamond can approach the missing timelike boundary without converging in the physical spacetime. Stable causality survives; global hyperbolicity does not.

The conclusion is exact but limited. AdS is not geodesically singular merely because its diagram has a finite edge, and deterministic evolution is not impossible. One must formulate an initial-boundary value problem or choose a self-adjoint evolution domain.

Boundary conditions complete the dynamics

Boundary Conditions and AdS Evolution uses the scalar asymptotic modes

$$\phi=z^{\Delta_-}\alpha+z^{\Delta_+}\beta+\cdots, \qquad \Delta_\pm=\frac d2\pm\sqrt{\frac{d^2}{4}+m^2\ell^2},$$

to compare Dirichlet-, Neumann-, mixed-, and transparent-type behavior. Which modes are admissible depends on the mass, dimension, energy norm, and BF bound. Reflecting conditions make symplectic flux vanish; transparent conditions require an open or coupled boundary system; mixed conditions must pass flux, well-posedness, and positivity tests.

Gravity has analogous but constrained data. Fixing a boundary conformal metric gives a Dirichlet-type phase space. Allowing that metric to vary introduces boundary dynamics and changes the symplectic form and symmetry group. Boundary terms and counterterms must match the same choice. No causal argument selects one universal condition.

Ishibashi and Wald’s self-adjoint-extension framework makes the field, dimension, mass, and energy dependencies explicit Ishibashi and Wald 2004, §2 and §§3.1–3.3, pp. 2985–3012.

Expansion, state data, and charges

Asymptotic Expansions near the AdS Boundary adopts Fefferman–Graham normal form. The boundary representative $g_{(0)}$ fixes lower even radial coefficients locally through the Einstein equation. At order $d$, the constraints leave transverse-traceless global data in $g_{(d)}$. Even boundary dimension generally introduces a $z^d\log z$ obstruction/anomaly term.

This split prevents two errors. A nonzero lower coefficient can reflect the curvature of $g_{(0)}$ rather than a physical excitation, while a black brane and Poincaré AdS can share flat $g_{(0)}$ but differ in $g_{(d)}$. The expansion is asymptotic and local; it does not reconstruct global interior topology or guarantee convergence.

Asymptotic Symmetries and Charges in AdS derives allowed boundary conformal transformations under standard Dirichlet data and compares Hamiltonian, covariant, electric-Weyl, and counterterm charges. Agreement requires the same dimension, boundary representative, orientation, field falloffs, counterterm scheme, and vacuum normalization. Exact AdS can have zero electric-Weyl energy yet a nonzero scheme-dependent Casimir energy; state differences still agree.

Causality, trapping, and open stability questions

AdS Causality, Trapping, and Stability Interfaces separates three mechanisms. Boundary reflection returns radiation to the bulk. Commensurate normal-mode frequencies create resonant channels. Stable trapping, including reflected rays confined between the boundary and an angular-momentum barrier, produces quasimodes and slow decay. Each makes dispersion-based nonlinear estimates harder, but none alone proves collapse.

As of 2026-08-12, the broad nonlinear stability question remains formulation dependent. Vacuum versus matter, dimension, symmetry, boundary condition, data topology, background, and norm change the claim. Numerical weak-turbulence evidence, resonant approximations, time-periodic solutions, linear decay theorems, and nonlinear formation results have different evidentiary force. New model-specific verdicts belong in dated research records.

Boundary causality is another distinct constraint: bulk propagation should not beat the causal relation of the fixed boundary conformal metric under appropriate equations and energy assumptions. It is neither global hyperbolicity nor a stability theorem.

Review the chapter

A scalar on global AdS has mass in a range where two normalizable asymptotic modes exist. A calculation fixes neither coefficient, evolves bulk initial data, and reports nonunique solutions. Is this evidence that determinism fails fundamentally, or that the problem is incomplete? What tests should a proposed boundary condition pass?

Solution

The problem is incomplete. Timelike infinity permits incoming data, so bulk initial data do not select a unique solution. One must choose a boundary domain: fix one mode, impose an admissible mixed relation, or couple to explicit boundary degrees of freedom.

The condition should make the symplectic/energy flux vanish for a closed system or balance it in the combined system; define a self-adjoint or well-posed evolution operator; satisfy corner compatibility with initial data; preserve the intended asymptotic symmetries; and possess an energy bound or other stability estimate. Passing flux alone does not prove positivity. Transparent absorption changes the system and cannot be compared directly with reflective AdS stability.

What this chapter enables

After this chapter, you should be able to classify AdS infinity invariantly, prove the compact-diamond failure, formulate why boundary data are required, derive the FG source/state split, compare charge schemes, and state stability claims with their boundary and data qualifications.

The next chapter asks how causal and energy assumptions restrict topology. Begin with Topology of Cauchy Surfaces.

References

• de Haro, Sebastián, Kostas Skenderis, and Sergey N. Solodukhin, “Holographic Reconstruction of Spacetime and Renormalization in the AdS/CFT Correspondence”, Communications in Mathematical Physics 217, §§2–4, pp. 595–622, 2001.
• Ishibashi, Akihiro, and Robert M. Wald, “Dynamics in Non-Globally-Hyperbolic Static Spacetimes III: Anti-de Sitter Spacetime”, Classical and Quantum Gravity 21, §2 and §§3.1–3.3, pp. 2981–3014, 2004.

Dated Research assessment. For open questions and recent results, follow the link below. Each assessment states the date through which it considers results and the limits of its conclusions.

• Global Dynamics of Asymptotically AdS Spacetimes