One-State Property of Closed Universes

• One-State Property of Closed Universes is defined by the collapse of all semiclassical states into a single, unique quantum state via non-perturbative gravitational effects.
• It is demonstrated in models such as JT gravity and holographic frameworks, where wormhole contributions lead to a rank-one inner product matrix.
• This unique state has profound implications for the quantum-to-classical transition, bulk quantum mechanics, and the interplay with baby universes.

A closed universe, defined as a spacetime configuration with no spatial boundary, exhibits a remarkable and highly constrained quantum structure in various regimes of quantum gravity. The “one-state property” asserts that after a complete non-perturbative treatment, the physical Hilbert space of closed universes is exactly one-dimensional: there exists a unique quantum state, and all distinct semiclassical states prepared by different boundary conditions become gauge-equivalent. This property emerges in models such as Jackiw–Teitelboim gravity, the Hartle–Hawking no-boundary proposal, and general holographic and algebraic AdS/CFT frameworks. The principle has profound implications for the quantum-to-classical transition, the role of baby universes, factorization, and the emergence of bulk quantum mechanics from a purely classical ensemble.

1. Model Realizations and Formulation

Closed universes have been explicitly studied in low-dimensional gravity models—most notably, Jackiw–Teitelboim (JT) gravity coupled to matter and associated topological toy models. The canonical JT action (in Euclidean signature, $8\pi G_N = 1$) is

$$I_{JT} = -\frac{1}{2} \int d^2x \sqrt{g}\; \phi (R+2) + I_{matter},$$

where $\phi$ is the dilaton enforcing constant negative curvature, and $I_{matter}$ can represent massive arrivals or operator insertions $O_i$ (Usatyuk et al., 2024). In all these cases, the closed universe preparation consists of a Euclidean path integral terminating on a compact slice (e.g., an $S^1$ of length $b$) and possible boundary insertions.

Toy models supplement this with a combinatorial description: each state is a collection of circles (with flavor indices), and the amplitude is obtained by summing over all two-manifolds whose boundaries match the circles, weighted by bulk and matter actions. The corresponding partition function or inner product between two boundary states $|\psi_i\rangle$ and $|\psi_j\rangle$ is

$$Z = \sum_{M,\; pairings} e^{-I(M)},\qquad \langle\psi_i|\psi_j\rangle \equiv Z_{ij},$$

where $I(M) = -S_0 \chi(M) + I_{matter}$, $\chi(M)$ the Euler characteristic (Usatyuk et al., 2024).

2. Semiclassical States and Perturbative Structure

At the semiclassical or perturbative level, closed universes appear to admit a rich structure of nearly orthogonal states. In JT+matter models with large matter mass $m \gg 1$, semiclassical states are constructed via operator insertions at different Euclidean times or by varying $\beta$, producing a family of cosmological wavefunctionals $|\psi_{\beta, i}\rangle$ (Usatyuk et al., 2024). Their overlaps, at leading order, form a Gram matrix:

$$\langle\psi_{\beta, i}|\psi_{\beta, j}\rangle = \delta_{ij} + O(e^{-2S_0}),$$

suggesting the apparent existence of many independent (labeled by flavor, geometry, and location of insertions) semiclassical configurations.

The Lorentzian continuation along an appropriate slice produces big-bang/big-crunch FRW universes, and the spectrum of states can appear unbounded at the perturbative (saddle-point) level. Similar behavior holds in topological models, where $k$ possible flavors suggest $k$ independent states prior to including non-perturbative effects.

3. Non-Perturbative Collapse to a Unique State

Non-perturbatively, wormhole topologies and genus-sum contributions restore a permutation symmetry across boundaries and enforce dramatic constraints. Explicit calculations of the inner product matrix $M_{ij} = \langle\psi_i|\psi_j\rangle$ show that all higher moments satisfy

$\mathrm{Tr}(M^n) = [\mathrm{Tr} M]^n,$

implying $M_{ij}$ is a rank-one matrix—so the Gram matrix only possesses a single nonzero eigenvalue, and all vectors are colinear (Usatyuk et al., 2024, Usatyuk et al., 2024). The result holds in JT gravity, in topological toy models, and for Euclidean path integrals with arbitrary boundary data (Abdalla et al., 2 Feb 2026). The general logic is:

• For any $n$-point correlation computed as a gravitational path integral, the ability to permute boundaries and glue together bras and kets leads to the functional form above.
• Analytic continuation to $n \to 0$ (replica trick) enforces that the closed-universe Hilbert space has rank one.

In the holographic framework, similar factorization conditions lead to every state being proportional to the Hartle–Hawking wavefunction:

$|\Psi\rangle = Z[\Sigma_B]\,|HH\rangle,$

and the space of states again collapses to one-dimensionality (Usatyuk et al., 2024).

A mathematically rigorous version appears in large $N$ AdS/CFT, where, under uniform convergence of single-trace correlators and energy boundedness, all possible closed-universe states converge to a unique pure state in the infinite-$N$ Hilbert space (Gesteau, 17 Sep 2025). The no-go theorem shows that, absent further modifications such as coarse-graining over heavy operators or ensemble-averaging, semiclassical baby universes do not emerge at the strict large-$N$ limit.

4. Hartle–Hawking State, Classical Probability, and Emergent Bulk Quantum Mechanics

The gravitational path integral for closed universes, notably in the Hartle-Hawking no-boundary proposal, returns only a single complex number, not a wave-functional over nontrivial configurations:

$$\Psi_{HH} = \int_{g\,|\,\partial M = \emptyset} Dg\,e^{-S[g]},$$

and the associated Hilbert space is thus one-dimensional (Zhao, 5 Feb 2026, Abdalla et al., 2 Feb 2026). Any classical probability structure—formally, an $L^2(\Omega, p(\alpha)\,d\alpha)$ space of α-sectors—carries no quantum noncommutative structure; operators act by multiplication and commute. The operator algebra is thus commutative and carries only classical statistics.

Nevertheless, bulk quantum mechanics for an observer within the universe emerges via “patch operators,” which select laboratory regions and insert operators there. In each α-sector, these patch operators define a linear functional, and the GNS construction then provides the usual (generally finite-dimensional) quantum Hilbert space for local observables (Zhao, 5 Feb 2026). Thus, the bulk quantum theory is reconstructed from classical α-ensemble data—an explicit realization of Wheeler’s “It from Bit.” The baby-universe Hilbert space remains purely classical; quantum structure is present only locally, from the viewpoint of internal observers.

In one-dimensional solvable models of de Sitter, this construction is manifest: the only global state is $|\Psi_{HH}\rangle$, while any patch operator $X$ is recovered as a (random) classical variable $O_X = \langle\Psi_{HH}| X |\Psi_{HH}\rangle$. Reconstruction of the full bulk quantum algebra on a $Z_L$-dimensional space is possible despite the baby-universe space's classical nature (Zhao, 5 Feb 2026).

5. Physical Interpretations, Implications, and Emergent Observers

The one-state property fundamentally reclassifies distinct semiclassical cosmological boundary conditions as gauge-equivalent when summed over all wormhole topologies. From the “outside,” there is precisely one quantum state for the entire closed universe. However, from the “inside”—that is, from the perspective of physical observers in a specific realization of the universe—nontrivial quantum mechanics can emerge from the structure of patch operators and the classical α-ensemble statistics (Zhao, 5 Feb 2026).

This dichotomy has far-reaching implications:

• Ensemble and α-sector interpretations: Each member of an α-sector ensemble has one state, but observable correlations and entropy (e.g., de Sitter entropy $S_{dS} = \log Z_L$) reflect coarse-grained information about underlying classical ensemble data (Zhao, 5 Feb 2026).
• Black hole evaporation parallels: A fully evaporated black hole interior—viewed as a closed universe entangled with radiation—illustrates that many apparent interior degrees of freedom can collapse non-isometrically to the unique exterior state (Usatyuk et al., 2024).
• Islands and mixed states: Entanglement with multiple non-gravitating reference systems can induce genuinely mixed states in the closed universe, with entropy bounded by half the coarse-grained total (see $S_{closed} \le \frac12 S_{coarse}$ in multiboundary wormholes) (Fallows et al., 2021). With only one external system, the encoding remains pure-state, exemplifying the severe restriction gravity imposes on the number of independent closed-universe states.

6. Factorization, Holography, and Large $N$ Tension

In holographic quantum gravity, factorization of partition functions and correlators for disconnected preparation boundaries ensures that all closed-universe states are proportional to the Hartle–Hawking state (Usatyuk et al., 2024):

$\Psi[Σ_1 \cup Σ_2] = \Psi[Σ_1]\,\Psi[Σ_2].$

The inner product matrix for all such states, $G_{ij}$, is of rank one, confirming the one-state property.

Large-$N$ holographic AdS/CFT frameworks sharpen this statement via no-go theorems: if single-trace correlators converge uniformly in the limit and support is restricted to $O(1)$ energy, a single closed-universe state arises at infinite $N$ (Gesteau, 17 Sep 2025). Evading this result (and allowing for a larger Hilbert space) requires either ensemble/averaged large-$N$ limits or coarse-graining over heavy operator data—modifications beyond the traditional strict large-$N$ or semiclassical path integral.

7. Refinements, Exceptions, and Future Directions

Several caveats and potential extensions exist:

• Algebraic holography and the averaged large-$N$ limit: When $N$-dependent oscillations are rapidly varying, pointwise large-$N$ limits fail, but an averaged limit recovers appropriate thermal and mixed correlation structure, yielding a larger Hilbert space in a macroscopic closed universe (Liu, 17 Sep 2025). This procedure is distinct from ensemble averaging over coupling constants.
• Final-state projection and EFT recovery: In the limit of vanishing bulk entanglement, holographic encoding maps collapse to a one-dimensional image; abandoning the final-state projection or coarse-graining over heavy operators restores the full effective-field-theory (EFT) Hilbert space (Antonini et al., 14 Jul 2025).
• Islands in multi-component environments: When closed universes are entangled with more than one non-gravitating system, the pure-state encoding breaks down, and true mixed states arise, although their entropy remains capped (Fallows et al., 2021).
• Open problems: The construction of observables and the physics as experienced by an observer in a one-state Hilbert space remain subtle. Ensemble averaging restores a larger Hilbert space but at the cost of factorization. The non-perturbative justification of these properties for higher-dimensional gravity is an active area of research.

• (Usatyuk et al., 2024) Closed universes in two dimensional gravity
• (Zhao, 5 Feb 2026) "It from Bit": The Hartle-Hawking state and quantum mechanics for de Sitter observers
• (Usatyuk et al., 2024) Closed universes, factorization, and ensemble averaging
• (Abdalla et al., 2 Feb 2026) Consistent Evaluation of the No-Boundary Proposal
• (Gesteau, 17 Sep 2025) A no-go theorem for large $N$ closed universes
• (Antonini et al., 14 Jul 2025) The Baby Universe is Fine and the CFT Knows It: On Holography for Closed Universes
• (Liu, 17 Sep 2025) Towards a holographic description of closed universes
• (Fallows et al., 2021) Islands and mixed states in closed universes