Interior and Causality Requirements

Exterior recovery and a smooth interior are compatible only if their operators belong to one consistent quantum description. It is not enough to say that different observers “see different things.” A proposal must specify the operator algebra, the state or code subspace on which it acts, the commutators and accuracy, which records one observer can causally compare, and what happens at the singularity or inner horizon. These data decide whether two descriptions are complementary, redundantly encoded, genuinely cloned, or contradictory.

Required background. Quantum Information for Black-Hole Physics supplies no cloning, purification, operator-algebra recovery, and error norms.

Helpful background. Black-Hole Thermodynamics and Semiclassical Gravity supplies horizon states and backreaction; Quantum Fields in Curved Spacetime supplies local mode algebras; Black-Hole Information supplies AMPS and reconstruction; Quantum Gravity supplies relational observables and microscopic interior questions.

Put all operators in one description

For an old evaporating hole, let $E$ denote early radiation, $b$ a late exterior mode, and $\widetilde b$ the mode that local effective field theory assigns as its interior partner. A candidate framework must answer:

$$\mathcal A_E,\ \mathcal A_b,\ \mathcal A_{\widetilde b} \subseteq \mathcal A_{\mathrm{phys}}\ ?$$

The question mark is essential. In a nongravitating factorized model these may be tensor factors. In gravity, dressing and constraints can give overlapping or noncommuting algebras, as the perturbative dressing analysis makes explicit Donnelly and Giddings 2016, §§2–5. If $\widetilde b$ is reconstructed from the radiation, the proposal must state whether it is the same logical operator represented in two physical regions or a second independent degree of freedom.

On a code subspace $\mathcal H_{\mathrm{code}}$ with encoder $V$, two physical representatives $O_E$ and $O_{\mathrm{bulk}}$ describe one logical operator $O$ when

$$O_E V\lvert\psi\rangle =O_{\mathrm{bulk}}V\lvert\psi\rangle =VO\lvert\psi\rangle$$

for every allowed $\lvert\psi\rangle$, or within a named norm and error. That is redundant encoding, not cloning. Outside the code subspace the equality may fail, so domain size and superpositions are part of the claim.

No cloning and causal access are distinct tests

The no-cloning theorem forbids a state-independent linear operation that makes two independent perfect copies of an arbitrary unknown state. That algebraic contradiction does not require the copies to meet an observer. Establishing that two semiclassical labels denote independent quantum copies is the crucial premise. An operational duplication experiment adds the requirement that one observer can access both records and bring the results into one causal comparison. The decisive inner-product argument is given in Wootters and Zurek 1982, complete linear proof.

An exact code gives a useful check. Suppose a full logical qubit has code-preserving representatives $X_E$ and $Z_I$ on two commuting physical regions, each allegedly carrying the complete qubit. Their reconstruction identities imply

$$[X_E,Z_I]V=V[X,Z].$$

The left side vanishes by physical commutativity, while the right side does not because logical Pauli $X$ and $Z$ do not commute. Complete exact qubit recovery from both independent regions is therefore impossible. A shared classical label, overlapping supports, or a restricted logical algebra does not make that stronger claim.

The AMPS setup sharpens the issue differently. Unitarity after the Page time requires the newly emitted mode $b$ to correlate with $E$ strongly enough that $S(Eb)\lt S(E)$, while smooth-horizon EFT makes $b\widetilde b$ nearly vacuum-pure. Strong subadditivity or monogamy then produces a contradiction under the common factorization assumptions. The logical chain and its mode qualifications are in Almheiri et al. 2013, §2 and Appendices A–C. The result identifies an inconsistent assumption set; it does not select a unique interior.

Causal accessibility and backreaction

Every duplication experiment needs worldlines and deadlines. The exterior observer must collect enough early and late radiation, perform any decoding, enter the hole, receive an interior signal, and compare records before the relevant singular boundary. Near a Schwarzschild horizon, the available Kruskal interval can make the interior signal energy exponentially large in the exterior waiting time. Scrambling-time retrieval and Page-time retrieval therefore lead to different causal and backreaction estimates.

The calculation must include detector localization, signal bandwidth, clock accuracy, gravitational dressing, and stress energy. If the measurement or message changes the geometry by order one, the original causal diagram cannot be used to claim a contradiction. Conversely, invoking backreaction without a quantitative threshold does not establish consistency. Harlow works through the complementarity, decoding, and causal-patch distinctions in Harlow 2016, §§5.6–5.7 and 7.1.

State dependence, linearity, and overlaps

State-dependent interior operators try to reproduce smooth correlators using an operator map tied to an equilibrium state or small algebra. Their test is not whether one reference state has the right two-point function. It is whether overlapping code subspaces assign compatible operators, whether superpositions obey the Born rule, and whether the construction preserves commutators and causal propagation to its stated accuracy.

Interior strategy	Algebraic requirement	Adverse check
Causal-patch complementarity	No single observer algebra contains two independent records	Construct the full retrieval-and-infall protocol and close its energy budget.
Error-correcting reconstruction	Multiple physical representatives act as one logical operator on a code	Enlarge the code and bound reconstruction error on superpositions.
State-dependent mirrors	State-labelled operator domains reproduce equilibrium correlators	Compare overlapping domains and Born probabilities for mixed superpositions.
Nonlocal transfer	Modified Hamiltonian carries correlations across a declared scale	Verify commutators, energy flux, signaling constraints, and exterior observables.
Horizon-scale microstructure	Conventional interior algebra is replaced by microscopic degrees of freedom	Predict both infalling response and exterior recovery for typical states.

Papadodimas and Raju give a concrete small-algebra mirror construction in Papadodimas and Raju 2014, §§2–5. Its restricted domain is part of the proposal, while global linearity and overlap objections remain contested. General quantum-information and explicit string-theory constructions are outside this page’s scope; the criterion here is consistency when those structures are used as a black-hole interior.

Singularities and inner horizons are separate tests

A proposal must also say where interior evolution ends. Removing an exterior duplication experiment does not resolve a spacelike singularity. Replacing a singularity by a bounce does not by itself stabilize a Cauchy horizon. Weak metric extendibility, bounded curvature, unitary relational evolution, and a smooth detector response are non-equivalent criteria.

The operator description should therefore include the domain of evolution, boundary data, conserved charges, and matching to the semiclassical exterior. If a final-state condition or postselection is imposed, its normalization, interaction sensitivity, linearity, and signaling consequences enter the same algebraic test. If no classical interior exists microscopically, the proposal must explain why effective interior correlators nevertheless emerge and where their approximation fails.

Results and open questions through 2026-08-19

Proposal or result	Assumptions	Controlled result	Limitation
AMPS incompatibility, Almheiri et al. 2013, §§2–3	Old hole, declared $E$–$b$–$\widetilde b$ factorization, sufficiently pure near-horizon pair, and fine-grained unitarity	Entropy inequalities make the assumptions mutually inconsistent	Does not prove a unique firewall or settle gravitational factorization
Causal complementarity, Susskind, Thorlacius, and Uglum 1993, §§2–4	Specified exterior waiting time, retrieval model, infalling trajectory, interior deadline, and signal-energy estimate	Can prevent one observer from comparing two records in the named regime	Depends on the decoder, geometry, detector localization, and backreaction
Entanglement-wedge recovery, Dong, Harlow, and Wall 2016, complete Letter and proof	Finite regulated code, declared logical algebra and boundary region, and relative-entropy hypotheses	Establishes one logical operator with multiple physical representatives on the code	Nonperturbative error, code size, efficiency, and generic evaporation remain conditional
State-dependent interiors, Papadodimas and Raju 2014, §§2–5	Equilibrium state, small algebra, restricted excitations, and state-labelled mirror prescription	Reproduces selected equilibrium interior correlators in the stated domain	Overlap, superposition, Born-rule, and enlarged-domain objections remain contested
Soft nonlocal transfer, Giddings 2013, §§2–5	Effective atmosphere-scale interaction with specified state, support, bandwidth, and rate	Gives a parameterized correlation-transfer channel	Microscopic origin, commutators, stress tensor, and causal signaling require independent control
Horizon-scale microstructure, Bena et al. 2022, §§2–8	Protected string-theory sectors and explicitly constructed state families or geometries	Supplies many nontrivial microscopic states and smooth horizon-scale solutions	Typical-state interior, generic collapse, and full radiation dynamics remain unresolved

Competing positions place the failed assumption in different columns: factorization, smoothness, state-independent linear operators, strict locality, or the existence of a conventional interior. The conclusions could change if a large code-subspace construction passes overlap and Born-rule checks with controlled error, if a causal duplication protocol closes its decoder and backreaction budget, or if an explicit counterexample shows two independently measurable copies in one physical observer algebra.

Check your understanding

An interior operator and a radiation operator have identical action on every state in a code subspace. Has the logical qubit been cloned?

Solution

Not necessarily. If both are physical representatives of the same logical operator, their agreement is redundant encoding. Cloning would require two independent subsystems that can be manipulated and measured as separate copies of an arbitrary unknown state. One must inspect the commutant, erased-region recovery, code projector, and causal accessibility. The conclusion can change outside the code subspace, so superpositions and the reconstruction error must also be tested.

References

• Almheiri, Ahmed, Donald Marolf, Joseph Polchinski, and James Sully, “Black Holes: Complementarity or Firewalls?” Journal of High Energy Physics 02 (2013) 062. arXiv
• Bena, Iosif, Emil J. Martinec, Samir D. Mathur, and Nicholas P. Warner, “Fuzzballs and Microstate Geometries: Black-Hole Structure in String Theory,” arXiv:2204.13113 (2022). Open manuscript
• Dong, Xi, Daniel Harlow, and Aron C. Wall, “Reconstruction of Bulk Operators within the Entanglement Wedge in Gauge-Gravity Duality,” Physical Review Letters 117 (2016) 021601. DOI. Open manuscript
• Donnelly, William and Steven B. Giddings, “Observables, Gravitational Dressing, and Obstructions to Locality and Subsystems,” Physical Review D 94 (2016) 104038. arXiv
• Giddings, Steven B., “Nonviolent Nonlocality,” Physical Review D 88 (2013) 064023. DOI. Open manuscript
• Harlow, Daniel, “Jerusalem Lectures on Black Holes and Quantum Information,” Reviews of Modern Physics 88 (2016) 015002. DOI
• Papadodimas, Kyriakos and Suvrat Raju, “State-Dependent Bulk-Boundary Maps and Black Hole Complementarity,” Physical Review D 89 (2014) 086010. arXiv
• Susskind, Leonard, Larus Thorlacius, and John Uglum, “The Stretched Horizon and Black Hole Complementarity,” Physical Review D 48 (1993) 3743–3761. DOI. Open manuscript
• Wootters, William K. and Wojciech H. Zurek, “A Single Quantum Cannot Be Cloned,” Nature 299 (1982) 802–803. DOI