de Sitter Geometry and Holographic Obstacles

De Sitter geometry has a smooth spacelike conformal boundary and observer-dependent cosmological horizons, not the timelike boundary of anti-de Sitter space. Those structural differences change the meaning of states, time evolution, observables, and reconstruction before any microscopic proposal is chosen. This page fixes $(d+1)$-dimensional de Sitter radius $L$, the Bunch–Davies state when a vacuum is needed, and either the global spacetime or one geodesic observer’s static patch. All holographic interpretations discussed here remain proposal-level as of 2026-08-20.

Required background. Gauge Theory and String-Theory Preparation provides constrained observables and analytic continuation.

Helpful background. Quantum Gravity frames diffeomorphism-invariant observables; Quantum Information supplies state-space and entropy distinctions; Causality and Global Structure provides conformal boundaries and horizons; and Semiclassical Gravity provides horizon temperature and entropy.

Global de Sitter and one observer’s patch

Embed de Sitter space as

$$-X_0^2+\sum_{i=1}^{d+1}X_i^2=L^2$$

in $(d+2)$-dimensional Minkowski space. Global coordinates $X_0=L\sinh(t/L)$ and $X_i=L\cosh(t/L)n_i$, with $\sum_i n_i^2=1$, give

$$ds^2=-dt^2+L^2\cosh^2(t/L)d\Omega_d^2.$$

Both $t\to-\infty$ and $t\to+\infty$ approach spacelike conformal spheres, $\mathcal I^-$ and $\mathcal I^+$. There is no spatial infinity and no de Sitter Killing field timelike everywhere. Witten 2001, pp. 1–4, equations (1)–(8) makes this geometric obstruction explicit and explains why a globally positive conserved energy is unavailable.

Choose the geodesic at the north pole. On its causal diamond, embedding coordinates

$$X_0=\sqrt{L^2-r^2}\sinh(t_s/L),\qquad X_{d+1}=\sqrt{L^2-r^2}\cosh(t_s/L)$$

produce the static metric

$$ds^2=-f(r)dt_s^2+\frac{dr^2}{f(r)}+r^2d\Omega_{d-1}^2, \qquad f(r)=1-\frac{r^2}{L^2}.$$

The Killing field $\partial_{t_s}$ is timelike only for $r\lt L$ and becomes null at the observer’s horizon. Thus static time is physically useful but patch-dependent.

Results and limitations through 2026-08-20

The formulation axes are dimension, global or static region, observer, vacuum, asymptotic conditions, gravitational dressing, and observable class.

Evidence or position	Controlled result	Limitation or adverse discriminator	Date
Global geometry and observability, Witten 2001, pp. 1–7, equations (1)–(13)	Spacelike $\mathcal I^\pm$, absence of a globally timelike generator, and an observer’s finite causal region follow geometrically	Proposed “meta-observables” between asymptotic boundaries need not be measurable by any in-spacetime observer	2001
dS/CFT proposal, Strominger 2001, §§2–8, pp. 4–14	In $\mathrm{dS}_3$, asymptotic symmetries and scalar correlators motivate Euclidean conformal data at spacelike infinity	Complex weights and possible nonunitarity obstruct a direct copy of a standard unitary AdS boundary CFT	2001
Horizon thermodynamics, Gibbons and Hawking 1977, pp. 2738–2741	Euclidean regularity assigns $T_{\mathrm{dS}}=1/(2\pi L)$ and $S_{\mathrm{dS}}=A/(4G)$ to the cosmological horizon	Thermodynamics alone does not identify a microscopic Hilbert space or prove that its dimension is $e^{S_{\mathrm{dS}}}$	1977
Observer algebra, Chandrasekaran, Longo, Penington, and Witten 2023, §§2–4, pp. 9–34	Adding an observer and imposing the gravitational constraint yields a type-II$_1$ static-patch algebra with a state-dependent entropy notion	The construction is semiclassical and algebraic; it is not a complete nonperturbative de Sitter dual	2023

The geometry and semiclassical temperature are controlled. Competing positions include wave-function data on $\mathcal I^+$, a microscopic theory for one static patch, observer complementarity, and an algebraic observable description. Claims that finite horizon entropy proves a finite-dimensional exact Hilbert space remain conjectural; causal-patch restriction by itself does not settle the state count.

Temperature and the causal diamond

Near $r=L$, set $r=L-\rho^2/(2L)$. After $t_s=-i\tau$, the leading two-dimensional metric is

$$ds_E^2\simeq d\rho^2+\frac{\rho^2}{L^2}d\tau^2.$$

Smoothness at $\rho=0$ requires $\tau\sim\tau+2\pi L$, hence

$$T_{\mathrm{dS}}=\frac{1}{2\pi L}.$$

This temperature uses static time normalized to proper time at $r=0$. A detector held at another fixed radius has proper time $d\tau_{\rm prop}=\sqrt{f(r)}\,dt_s$ and measures the Tolman temperature $T_{\rm loc}(r)=1/[2\pi L\sqrt{f(r)}]$; such a detector is accelerated. The horizon derivation does not assume a boundary dual. A static observer can receive signals only from the causal past of the future endpoint and influence only the causal future of the past endpoint. Their intersection is the static diamond. Global field operators outside it are not operational observables for that observer unless a proposal supplies a nonlocal relational construction.

The area $A=L^{d-1}\Omega_{d-1}$ gives the semiclassical entropy $A/(4G)$. It is robust as a saddle-point thermodynamic quantity. Interpreting it as the logarithm of a finite number of exact states, as entanglement across the horizon, or as an entropy of an observer algebra are inequivalent claims and require different regulators and ensembles.

Why the AdS reconstruction step fails

The usual AdS argument evolves timelike-boundary data with a boundary Hamiltonian and reconstructs a bulk domain from that data. In de Sitter, $\mathcal I^+$ is spacelike, not a timelike cylinder; specifying a future wave functional is not continuous-time access to an observable algebra. A single observer also cannot collect all of $\mathcal I^+$, and a generic perturbation may change whether a smooth future boundary exists.

Analytic continuation can relate individual saddle calculations, but it does not automatically continue the Hilbert-space inner product, integration contour, or observable interpretation. This is the precise failure of a direct dictionary transplant. It does not rule out holography; it demands that each proposal state what replaces boundary time, which region it describes, and how probabilities are extracted. A complete construction satisfying those demands remains unresolved.

What could change these conclusions

The conclusions could change if a theorem constructs a nonperturbative state space and observable algebra for specified de Sitter boundary conditions, if an independent reconstruction calculation maps a declared boundary data class into a static diamond, or if a counterexample invalidates the assumed future boundary. The conclusions also depend on whether a new no-go result, operational protocol, or controlled string construction changes the relation between global and observer-patch descriptions.

Exercises

Use the embedding formulas to identify the static horizon, derive its temperature, and explain which assumption in an AdS-style boundary evolution argument fails.

Solution

The static embedding covers $X_{d+1}^2-X_0^2=L^2-r^2\geq0$, so its Killing field becomes null at $r=L$. Expanding the Euclidean metric there gives a plane in polar variables only when $\tau$ has period $2\pi L$; therefore $T=1/(2\pi L)$. The reconstruction comparison fails because future conformal infinity is spacelike and is not a timelike boundary carrying a globally positive Hamiltonian. Moreover, one observer sees only a causal diamond. These geometric facts do not disprove holography, but they require a different observable and state-space formulation.

References

• Chandrasekaran, Venkatesa, Roberto Longo, Geoff Penington, and Edward Witten. “An Algebra of Observables for de Sitter Space.” Journal of High Energy Physics 02 (2023) 082. arXiv:2206.10780.
• Gibbons, G. W., and S. W. Hawking. “Cosmological Event Horizons, Thermodynamics, and Particle Creation.” Physical Review D 15 (1977) 2738–2751. DOI: 10.1103/PhysRevD.15.2738.
• Strominger, Andrew. “The dS/CFT Correspondence.” Journal of High Energy Physics 10 (2001) 034. arXiv:hep-th/0106113.
• Witten, Edward. “Quantum Gravity in de Sitter Space.” In Strings 2001. arXiv:hep-th/0106109.