Microstates, Typicality, and Coarse Graining

Black-hole entropy counts states only after a sector has been defined, and a large state count does not make every state geometrical or every observable thermal. This chapter separates four claims: a microcanonical subspace has a large dimension; most vectors in that subspace look thermal to a restricted subsystem; individual energy eigenstates may satisfy eigenstate thermalization for selected simple operators; and coarse graining may admit a common semiclassical geometry. Each claim uses a different measure, observable algebra, and approximation. Their careful combination clarifies why microscopic purity can coexist with robust macroscopic black-hole behavior.

Helpful background. Semiclassical Gravity supplies Bekenstein–Hawking entropy and the thermodynamic variables used to label a sector; Quantum Information supplies reduced states, trace distance, and concentration of measure; and Quantum Fields in Curved Spacetime supplies the simple exterior observables whose limited resolving power is being tested.

State spaces, measures, and observable algebras

Let $P_{E,\Delta E,Q}$ project onto states satisfying an energy window and a set of conserved charges $Q$. The microcanonical state is

$$\Omega_{E,\Delta E,Q} =\frac{P_{E,\Delta E,Q}}{d_{E,\Delta E,Q}}, \qquad d_{E,\Delta E,Q}=\operatorname{Tr}P_{E,\Delta E,Q},$$

and its entropy is $S_{\rm micro}=\log d_{E,\Delta E,Q}$. This equation defines a sector; it does not say that every vector in it has a smooth interior.

Typicality adds a probability measure, usually Haar measure on that constrained subspace. ETH adds a Hamiltonian and a declared class of operators. Geometric coarse graining adds a map from microscopic states or expectation values to effective fields and discards unresolved data. A conclusion can pass one of these tests and fail another. For example, an integrable system has a large microcanonical state space but can violate standard ETH through additional conserved quantities.

Route through the chapter

The pages follow that logical order.

1. Microcanonical Black-Hole State Spaces fixes energy windows, charges, regulators, and the logarithmic dimension compared with black-hole entropy.
2. Canonical Typicality proves concentration of a small subsystem’s reduced state around the constrained ensemble and states the effective-environment bound.
3. Eigenstate Thermalization and Simple Observables identifies the observable class, energy window, symmetry sector, and finite-size tests required by ETH.
4. Coarse-Grained Geometries asks which averaging or restriction makes many exact states share effective geometric data.
5. Microstate Geometries in String Theory evaluates whether known smooth solutions sample typical states and sends their detailed construction to the dedicated string/holography treatment.
6. Entropy, Degeneracy, and Observable Distinguishability derives the energy resolution, copy count, and operator precision needed to distinguish exponentially many nearby states.

Canonical typicality is a kinematic concentration result: it does not require the state to evolve toward equilibrium. Goldstein and collaborators state the constrained-subspace result and its dimension dependence in Goldstein et al. 2006, main theorem and concentration estimate. ETH is a different, dynamical statement about matrix elements in an energy basis; its standard ansatz, tests, and exceptions are reviewed in D’Alessio et al. 2016, §§ 2–4 and 6.

Results and open questions through 2026-08-19

This discussion reflects results available through 2026-08-19. The matrix distinguishes exact finite-dimensional statements from dynamical evidence and from microscopic gravity interpretations.

Framework and evidence	Result	Scope or limitation
Constrained Haar-typical states in Popescu, Short, and Winter 2006, Theorem 1	A small subsystem is overwhelmingly likely to be close to the canonical reduced state when its effective environment is large.	The theorem assumes a subsystem factorization and measure; it does not establish equilibration or gravity.
ETH ansatz and finite-system studies in D’Alessio et al. 2016, §§ 2–4 and 6	Chaotic nonintegrable models often show smooth diagonal matrix elements and entropy-suppressed off-diagonal elements for simple observables.	Symmetry sectors, energy windows, integrability, localization, and finite-size scaling can invalidate the naive ansatz.
Semiclassical entropy and common exterior data	Many microstates can agree on mass, charge, and low-point exterior observables to perturbative accuracy.	Equality of restricted data does not prove identical interiors or a common exact metric operator.
Explicit smooth microstate sectors in Bena et al. 2022, §§ 2–8	String constructions exhibit families of horizon-scale structure with controlled charges in special supersymmetric or near-supersymmetric regimes.	Known solutions need not form a typical basis for the full entropy-sized sector, and their construction is model dependent.

The permissive position combines typicality, ETH, and explicit constructions to argue that a shared coarse geometry is the natural description of nearly all simple measurements. The conservative position accepts restricted thermal behavior but withholds any conclusion about a typical state’s exact interior until the coarse-graining map and microscopic measure are derived. The evidence supports the restricted-observable statement most strongly; it remains unresolved whether a common semiclassical interior represents typical exact states in a single theory.

Independent checks and boundaries

Three checks prevent category errors. First, shrink the environment until its effective dimension is comparable to the subsystem: the typicality bound must lose power. Second, choose an integrable Hamiltonian or mix symmetry sectors: standard ETH can fail even though the state count remains exponential. Third, enlarge the observable algebra until it resolves individual level spacings: microstates that looked identical to simple probes become distinguishable.

This chapter does not redo stable black-hole thermodynamics, generic quantum-information theory, or detailed string and holographic construction. Those subjects are separate; the present scope is the chain of assumptions from counted state sector to restricted observable and effective geometry. Exact degeneracy, exponentially small splitting, and operational indistinguishability are kept distinct throughout.

The conclusions could change after a single-theory microscopic construction samples an entropy-sized typical sector, or a theorem extends canonical typicality to the relevant gravitational algebra without a hidden factorization assumption, or controlled finite-size data overturn the claimed ETH scaling for the black-hole observable class.

Check your understanding

A constrained subspace has dimension $10^{100}$, but the proposed environment for a subsystem has effective dimension two. Does the large total state count alone guarantee that the subsystem is close to its canonical state?

Solution

No. Canonical typicality is controlled by the effective environment dimension relative to the subsystem and by the chosen constrained measure, not by the total count alone. If the relevant environment has effective dimension two, the concentration bound can be weak and atypical reduced states need not be rare. One must calculate the reduced canonical state and the actual dimension ratio before making a thermal claim.

References

• Bena, Iosif, Emil J. Martinec, Samir D. Mathur, and Nicholas P. Warner. 2022. “Fuzzballs and Microstate Geometries: Black-Hole Structure in String Theory.” arXiv:2204.13113.
• D’Alessio, Luca, Yariv Kafri, Anatoli Polkovnikov, and Marcos Rigol. 2016. “From Quantum Chaos and Eigenstate Thermalization to Statistical Mechanics and Thermodynamics.” Advances in Physics 65, 239–362. arXiv:1509.06411.
• Goldstein, Sheldon, Joel L. Lebowitz, Roderich Tumulka, and Nino Zanghì. 2006. “Canonical Typicality.” Physical Review Letters 96, 050403. arXiv:cond-mat/0511091.
• Popescu, Sandu, Anthony J. Short, and Andreas Winter. 2006. “Entanglement and the Foundations of Statistical Mechanics.” Nature Physics 2, 754–758. arXiv:quant-ph/0511225.