Holographic Reconstruction and Gravitational Path Integrals
Holographic reconstruction and gravitational path integrals are linked by the AdS/CFT dictionary, but they have different logical roles. Reconstruction maps a declared bulk effective algebra into boundary observables, usually within a code subspace. A gravitational path integral estimates boundary quantities by a saddle expansion over metrics and topologies. Both are powerful in the semiclassical large-N regime; neither presently supplies generic finite-N local bulk operators or a nonperturbative definition of every gravitational saddle sum.
Evidence cutoff. 11 August 2026.
Required background. Holographic quantum error correction supplies the subregion framework; Euclidean gravitational path integrals, topology sums, and boundary conditions supplies the integration problem. Helpful background. Replica wormholes and saddle competition supplies entropy saddles, factorization and ensembles supplies a central obstruction, and approximate finite-N recovery supplies quantitative claim norms.
Dictionaries, code subspaces, and saddle sums
Section titled “Dictionaries, code subspaces, and saddle sums”Target classes include perturbative bulk fields and correlators, subregion reconstruction, entanglement wedges, black-hole interiors, entropy and Page curves, partition functions, spectral statistics, and topology-changing contributions. Every use must state the boundary theory or ensemble, asymptotic boundary conditions, large-N and large-gap assumptions, bulk state/code subspace, gravitational dressing, contour, and included saddle classes.
| Method | Applicability and input | Error and claim ceiling |
|---|---|---|
| HKLL-type smearing | Perturbative bulk field on a fixed asymptotically AdS background with known boundary correlators | , , interactions, dressing, horizons, and state dependence |
| Quantum-error-correction / entanglement-wedge reconstruction | Code subspace, bulk–boundary relative-entropy or algebraic equality | Approximate recovery norm, code size, center/area operator, finite-N leakage |
| Boundary modular-flow reconstruction | Boundary modular operator and bulk modular relation | Modular domains, code-subspace equality, practical nonlocality |
| Euclidean saddle expansion | Boundary conditions, action, measure, contour, counterterms, saddle set | Conformal-factor problem, negative modes, loop expansion, dominance, topology completeness |
| Replica/cosmic-brane and wormhole saddles | Analytic continuation in replica number and saddle competition | Continuation, replica symmetry, ensemble interpretation, factorization |
Hamilton, Kabat, Lifschytz, and Lowe reconstruct free bulk fields from smeared boundary operators in the semiclassical dictionary (Hamilton et al. 2006). Interactions require multitrace corrections, and gravitationally dressed observables are not exactly local. The apparent conflict between multiple boundary reconstructions is organized by a code subspace rather than an exact tensor factorization of the full Hilbert space.
Almheiri, Dong, and Harlow made the quantum-error-correction structure explicit (Almheiri, Dong, and Harlow 2015). JLMS related boundary and bulk relative entropies at leading semiclassical orders (Jafferis et al. 2016), and Dong, Harlow, and Wall derived entanglement-wedge reconstruction from the corresponding algebraic conditions (Dong, Harlow, and Wall 2016). The chain is conditional: approximate entropy equality yields approximate recovery only in a stated norm and state set; it does not create exact finite-N bulk locality.
Gravitational replica saddles reproduce generalized entropy and, in effective black-hole models coupled to baths, generate Page-curve behavior. Replica wormholes show that the semiclassical saddle catalog can change with replica number. They establish a gravitational computation under the chosen path integral and boundary conditions. Whether that path integral computes a single boundary theory, an ensemble average, or a coarse-grained quantity is a separate factorization question. The exact matrix-integral dual of JT gravity makes the ensemble interpretation demonstrable in that model (Saad, Shenker, and Stanford 2019); it cannot be transferred automatically to a fixed higher-dimensional CFT.
Error budgets, benchmarks, and independence
Section titled “Error budgets, benchmarks, and independence”Reconstruction errors should separate bulk EFT loops, finite coupling, finite N, code-subspace truncation, gravitational dressing, boundary smearing, and the operator norm or state-dependent metric used. Path-integral errors should separate gauge fixing and ghosts, measure/counterterms, loop determinant, negative modes, saddle omission, Stokes phenomena, topology sum, replica continuation, and nonperturbative corrections. Writing without a norm and state class is not an error bound.
Benchmarks include free bulk fields with exact boundary correlators, three-point Witten diagrams, Ryu–Takayanagi/JLMS relative entropy, tensor-network codes, AdS/CFT sectors with enhanced control, JT gravity and its matrix integral, and direct CFT spectra or correlators. HKLL, Witten diagrams, and a bulk effective action share the same dictionary and are correlated. Stronger independence comes from exact boundary CFT data, localization, bootstrap, lattice/statistical-model constructions, or a distinct nonperturbative dual not used to choose the bulk saddles.
Failure modes and alternatives
Section titled “Failure modes and alternatives”Reconstruction can fail outside the code subspace, across state-dependent entanglement-wedge transitions, or when gravitational dressing reaches the nominal complement. A saddle can be complex, possess a negative mode, or be subdominant; a topology sum can diverge. Replica symmetry may break. Wormholes connecting disconnected boundaries conflict with exact factorization in a fixed boundary theory unless an ensemble, alpha-state, or other mechanism accounts for them. Marolf’s analysis of ensemble interpretations also shows that positivity and factorization of putative ensemble elements require care (Marolf 2024).
Alternatives are not mutually exclusive: direct boundary reconstruction, conformal bootstrap, tensor networks, canonical quantization, Lorentzian path integrals, and algebraic QFT can constrain the same claim. A semiclassical saddle should be treated as evidence proportional to the independent boundary checks it survives.
Claim ceiling. In a large-N, semiclassical code subspace, bulk correlators and entanglement-wedge reconstruction can be controlled order by order. Replica saddles can establish generalized-entropy predictions in specified effective models. Generic finite-N bulk locality, a complete gravitational integration cycle, and single-theory factorization remain open.
Choosing a method
Section titled “Choosing a method”| Research question | Preferred route | Required guardrail |
|---|---|---|
| Perturbative local bulk correlator | HKLL plus interacting bulk EFT | Dressing, boundary conditions, and coupling order |
| Which boundary region reconstructs an algebra? | Entanglement-wedge/operator-algebra QEC | Code subspace and approximate recovery norm |
| Entropy or Page-curve saddle | Replica/cosmic-brane path integral | Replica continuation, competing saddles, ensemble/factorization statement |
| Nonperturbative spectrum or finite-N locality | Direct boundary CFT or exact dual first | Do not infer it from a semiclassical saddle alone |
| Disconnected boundaries or wormholes | Factorization test before interpretation | Identify single theory, ensemble, alpha sector, or inconsistency |
Related routes include holography and quantum gravity, replica, modular, and operator-algebra methods, islands and unitary evaporation, and bulk reconstruction beyond code subspaces. Reconstruction error norms provides optional depth.
Evidence cutoff and change criteria
Section titled “Evidence cutoff and change criteria”The finite set covers HKLL, quantum-error-correction and relative-entropy reconstruction, replica wormholes, exact JT ensemble structure, and factorization qualifications. Targeted arXiv, journal, INSPIRE, and citation-chain searches covered public evidence through 11 August 2026. Reassessment is triggered by a finite-N recovery theorem or counterexample, a corrected saddle/contour, an exact single-CFT derivation of a wormhole contribution, a factorization failure, or a nonperturbative definition of the path integral.
References
Section titled “References”- Almheiri, Ahmed, Xi Dong, and Daniel Harlow. “Bulk Locality and Quantum Error Correction in AdS/CFT.” Journal of High Energy Physics 2015, no. 4 (2015): 163. DOI.
- 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. arXiv.
- Hamilton, Alex, Daniel Kabat, Gilad Lifschytz, and David A. Lowe. “Local Bulk Operators in AdS/CFT: A Boundary View of Horizons and Locality.” Physical Review D 74 (2006): 066009. DOI.
- Jafferis, Daniel L., Aitor Lewkowycz, Juan Maldacena, and S. Josephine Suh. “Relative Entropy Equals Bulk Relative Entropy.” Journal of High Energy Physics 2016, no. 6 (2016): 004. DOI.
- Marolf, Donald. “On the Nature of Ensembles from Gravitational Path Integrals.” arXiv:2407.04625 (2024). arXiv.
- Saad, Phil, Stephen H. Shenker, and Douglas Stanford. “JT Gravity as a Matrix Integral.” arXiv:1903.11115 (2019). arXiv.