{"id":"29c49687-e8d6-441e-b31c-c8a33913e0d2","arxiv_id":"2606.18639","paper_version":3,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"When the RT area is O(1/G) and bulk entropy O(1), CCKLP-style reconstruction errors are exponentially small in G relative to area-function corrections.","lead":"Witten shows that in a generic small perturbation of exact entanglement wedge reconstruction, reconstruction errors are exponentially small in 1/G compared with state-dependent corrections to the RT area term. This quantifies how semiclassical gravity can have perturbative backreaction while reconstruction remains nonperturbatively accurate.","discovery_kind":"extension","skeptic_critique":{"model":"grok-4.5","headline":"No significant objection identified","rationale":"The paper's central claim is a quantitative hierarchy inside an existing toy model, not a derivation from first principles of holography. Once one accepts the CCKLP encoding (2.9) and the modeling assumption that low-energy eigenvalues are not exponentially small, the algebra that produces E[D_A1]/E[D_A1A2]≈1/d2^{2} is elementary and free of hidden steps. The reader's weakest-assumption remark correctly identifies the only place where the hierarchy could fail, but that place is already delimited by the author; it does not undermine the calculation that is actually performed. No independent technical gap (sign error, unjustified interchange of limits, missing term of the same order, etc.) was found. The verdict therefore remains ACCEPT with high confidence.","tokens_in":18978,"tokens_out":521,"duration_ms":4907,"concrete_test":"Recompute the ratio E[D_A1]/E[D_A1A2] for a pure bulk state whose reduced density matrix σ_A1 has one eigenvalue set to exp(-S_χ) while the rest are O(1/d1); if the ratio remains O(1/d2^{2}) rather than O(1), the hierarchy is robust even outside the paper's stated regime.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The reader's weakest-assumption note (GUE treating all modes equally, including those with exponentially small eigenvalues of σ_A1) is already flagged by the paper itself in §3.1 and after (3.18). Within the stated regime of the CCKLP model the paper simply assumes those eigenvalues are not exponentially small, so that Ξ remains O(d1^{2}) and the ratio E[D_A1]/E[D_A1A2]≈1/d2^{2} holds (eqs. 3.23, 3.26, 3.44). That is a modeling choice, not an internal inconsistency; the hierarchy is derived cleanly once the assumption is granted. No other soft spot in the central claim appears: the second-order Kubo-Mori expansion, the GUE second-moment calculation, the independence from recovery optimization (§3.9), and the mixed-state generalization (§3.10) are elementary and transparent.","agreement_with_reader":"agree"},"referee_report":{"model":"grok-4.5","summary":"The paper quantifies a feature of the CCKLP model of approximate entanglement wedge reconstruction. Starting from a generic GUE perturbation of an exact encoding map, it computes the second-order relative entropies D_A1 (reconstruction error on the low-energy factor) and D_A1A2 (distinguishability of the full A1A2 states) via the Kubo-Mori formula and the GUE second moment. When the high-energy dimension d2 is exponentially large (modeling an O(1/G) area term) while the low-energy dimension remains O(1), the ratio of ensemble averages is E[D_A1]/E[D_A1A2] ≈ 1/d2^{2} ~ exp(-c/G). The same hierarchy holds for a general high-energy density matrix (by Jensen convexity of L(x)=x coth(x/2)), after optimization over local recovery unitaries (by a Pinsker bound), and for mixed bulk states. Consequently the state dependence of the CCKLP area function is perturbative while reconstruction errors remain non-perturbatively small, matching semiclassical expectations. The paper also discusses why an area function rather than a central area operator is natural once continuum UV issues are taken into account.","tokens_in":19218,"tokens_out":776,"duration_ms":7638,"significance":"The result supplies a clean, quantitative bridge between the CCKLP construction and the expected separation of scales in holography: gravitational back-reaction is O(1) while reconstruction remains exponentially accurate. The derivation is elementary, fully explicit, and free of free parameters once the CCKLP ensemble and the invertibility assumption on the low-energy density matrix are granted. The extensions to non-maximally mixed high-energy states, recovery optimization, and mixed bulk states strengthen the claim without altering the leading exponential. The discussion of area functions versus area operators usefully situates the model relative to continuum QFT and to the exact-reconstruction alternative of Harlow. The note is short, self-contained, and of clear interest to the holographic quantum-information community.","major_comments":[],"minor_comments":[{"comment":"In the abstract and introduction the phrase \"exponentially small (in G) relative to corrections to the area function\" is slightly ambiguous; a parenthetical reference to the ratio ~1/d2^{2} would make the claim sharper for a reader who has not yet reached §3.","section":null},{"comment":"Section 3.1 and the paragraph after (3.18) correctly flag that exponentially small eigenvalues of σ_A1 would invalidate the O(d1^{2}) estimate of Ξ. A single additional sentence noting that a refined ensemble suppressing those modes would be needed for a full QFT treatment would make the modeling boundary even clearer.","section":null},{"comment":"The footnote acknowledging Claude Opus 4.8 is unusual for a journal article; if retained, it should be moved to the acknowledgements and phrased more formally.","section":null},{"comment":"A few typographical inconsistencies appear (\"W edge\", \"T echnical\", occasional missing spaces around math). A light copy-edit pass would remove them.","section":null}],"recommendation":"accept","confidential_remarks":"The manuscript is a short, high-quality note that cleanly settles one quantitative question left open by CCKLP. It is well within the scope of a serious hep-th journal and requires no further technical work. The only soft modeling assumption is already stated by the author; I see no reason to delay publication."},"author_rebuttal":null,"desk_editor":{"model":"grok-4.5","letter":"Witten takes the CCKLP setup (GUE perturbation of exact entanglement-wedge encoding, area defined as the residual entropy after optimal recovery) and computes the second-order relative entropies that control reconstruction error versus geometric backreaction. The punchline is quantitative: when the high-energy factor has dimension d2 ~ exp(c/G), the ratio of those relative entropies is ~1/d2^{2}, so reconstruction corrections are non-perturbatively small while the area function still feels O(1) state dependence. That hierarchy is what one expects from semiclassical gravity, and it is new; CCKLP did not evaluate it.\n\nThe calculation is elementary and carefully written. Kubo-Mori second-order relative entropy, the GUE second-moment formula, partial-trace moments, the d2 rescaling of the weights, Jensen convexity for non-maximally mixed χ, a Pinsker bound showing that recovery optimization cannot spoil the hierarchy, and the mixed-state generalization are all spelled out. Self-averaging estimates are given. The paper is also clear about its modeling assumptions: eigenvalues of the low-energy density matrix are taken not to be exponentially small (so that Ξ stays O(d1^{2})), and the GUE treats all modes equally. Those are limitations of the toy model, already flagged in §3.1, not hidden flaws in the algebra.\n\nThe discussion of area function versus area operator is short and sensible; it does not claim more than the calculation supports. Citations are appropriate and light.\n\nThis is a short technical note for people already working on holographic quantum error correction and the RT formula. It does not invent a new framework, but it sharpens an expected scale separation inside an existing one. The math is transparent enough that a referee can check it in an afternoon. I would send it out for peer review and would cite the ratio result if I were writing on approximate entanglement-wedge reconstruction.","headline":"Clean second-order calculation showing reconstruction error is exp-small relative to area-function backreaction in the CCKLP GUE model.","tokens_in":19776,"tokens_out":489,"would_cite":true,"duration_ms":5403,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"grok-4.5","headline":"In a generic perturbation of holographic codes, reconstruction errors are exponentially smaller than the back-reaction of bulk states on the area function.","keywords":["entanglement wedge reconstruction","Ryu-Takayanagi formula","area function","relative entropy","Gaussian unitary ensemble","holographic codes","gravitational back-reaction"],"falsifier":"Compute the same second-order relative entropies for a modified ensemble that suppresses modes whose eigenvalues are smaller than exp(-1/G); if the ratio D_A1 / D_A1A2 is no longer ~1/d_{2}^{2}, the claimed hierarchy fails inside the model.","tokens_in":19904,"feed_emoji":"∞","tokens_out":719,"duration_ms":5408,"temperature":0.7,"pith_summary":"Exact entanglement-wedge reconstruction forces the area term in the Ryu-Takayanagi formula to be a pure number, independent of the bulk quantum state. That would mean bulk matter never back-reacts on geometry. A recent model replaces exact reconstruction by a small generic unitary perturbation and defines an effective area function that does depend on the bulk state. This note shows that, once the area term is taken to be order 1/G and the bulk entropy order 1, the relative entropy that measures reconstruction error is smaller than the relative entropy that measures area-function corrections by a factor of order 1/d_{2}^{2}, where d_{2} is the dimension of the high-energy Hilbert space. That ratio is exponentially small in G. The calculation therefore places a quantitative hierarchy between the two corrections that matches the expectation from semiclassical gravity: back-reaction is perturbative while reconstruction remains exact to all orders in G. The same hierarchy continues to hold when the high-energy state is not maximally mixed and when the bulk state is mixed, and is essentially unaffected by the optimization over local recovery unitaries.","feed_headline":"Reconstruction errors exponentially smaller than geometry back-reaction","feed_subtitle":"A generic holographic-code perturbation keeps reconstruction exact to all orders in G while area depends on the bulk state.","key_machinery":"The Kubo-Mori second-order formula for relative entropy, evaluated on the second moments of a Gaussian unitary ensemble after partial traces over the high-energy factors. The ratio of the two resulting relative entropies is controlled by the dimensional factor d_{2}^{2} that arises from the rescaling of the Kubo-Mori weights when the high-energy density matrix is traced out.","core_discovery":"When the encoding map of a holographic code is deformed by a small Gaussian unitary perturbation, the second-order relative entropy that quantifies reconstruction error on the low-energy subspace is smaller than the relative entropy that quantifies state-dependence of the geometric entropy by a factor approximately equal to 1/d_{2}^{2}, with d_{2} the dimension of the short-distance Hilbert space. Because d_{2} scales as exp(c/G), reconstruction corrections are non-perturbatively small relative to gravitational back-reaction.","pith_inferences":[],"forward_implications":[],"fun_headline_variants":["Reconstruction errors exp-small vs geometry back-reaction","Holographic recon corrections non-perturbatively smaller than area shifts","Entanglement wedge errors ~1/d2^{2} times state-dependent area","Code perturbation: recon exact to all G orders; area state-dependent","Gaussian unitary deform: recon error exponentially tinier than geometry"],"cache_read_input_tokens":12160,"weakest_assumption_plain":"The random unitary is drawn uniformly from the full Hilbert space, treating every mode equally even when the low-energy density matrix has exponentially small eigenvalues.","fun_headline_variants_meta":{"raw":{"variants":["Reconstruction errors exp-small vs geometry back-reaction","Holographic recon corrections non-perturbatively smaller than area shifts","Entanglement wedge errors ~1/d2^{2} times state-dependent area","Code perturbation: recon exact to all G orders; area state-dependent","Gaussian unitary deform: recon error exponentially tinier than geometry"]},"model":"grok-4.5","effort":"low","cost_usd":0.002938,"raw_usage":{"total_tokens":1049,"prompt_tokens":743,"num_sources_used":0,"completion_tokens":75,"cost_in_usd_ticks":29380000,"prompt_tokens_details":{"text_tokens":743,"audio_tokens":0,"image_tokens":0,"cached_tokens":256},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":231,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":743,"tokens_out":75,"duration_ms":4544,"temperature":1.0,"reasoning_tokens":231,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-13T07:26:42.401528+00:00","model_set":{"reader":"grok-4.5"},"falsifier":"Compute the same second-order relative entropies for a modified ensemble that suppresses modes whose eigenvalues are smaller than exp(-1/G); if the ratio D_A1 / D_A1A2 is no longer ~1/d_{2}^{2}, the claimed hierarchy fails inside the model.","supporting_citations":[],"review_version":3}