{"id":"b6a296bf-e7dd-4931-b232-ec92562ab869","arxiv_id":"2508.17139","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"The stress-tensor part of thermal correlators develops complex-time singularities at exactly the locations where bulk geodesics reflect off the black hole singularity, and in Gauss-Bonnet holography ANEC saturation makes near-lightcone TT correlators vacuum-like.","lead":"This thesis shows that the stress-tensor sector of thermal correlators in holographic CFTs contains singularities matching geodesics that bounce off a black hole singularity, and that near-lightcone stress-tensor correlators become temperature-independent when the averaged null energy condition is saturated. It gives physicists a boundary-side probe of black hole interiors and a new link between causality bounds and correlation functions.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The claimed pole at τ_c is produced by a numerically fitted large-n OPE form with no quoted error control; a subleading 1/n term would change the exponent or location, so the geodesic match is not yet robust.","rationale":"The reader's weakest_assumption identifies both the near-boundary reliability and the numerically fitted large-n coefficients. I agree both are load-bearing, but I rate the large-n fit as the sharper and more directly testable point: the pole formula (3.60) is literally the integral of the fitted Λ_n^a, so any error in the fit enters the claimed exponent and location. The near-boundary assumption is also serious—if the series misses non-perturbative interior data, the extracted Λ_n are not the physical OPE coefficients—but it is explicitly flagged and would require a more specific failure mode. The proposed high-order recurrence test is feasible and would settle the fit issue without needing to solve the interior problem. Given the plausible independent geodesic and Ward-identity cross-checks, I do not see a reason to change the reader's CONDITIONAL verdict; the claim is attractive but not fully established until the asymptotic fit is brought under analytic or high-precision numerical control.","tokens_in":61619,"tokens_out":7351,"duration_ms":90070,"concrete_test":"Generate Λ_n to n=500 or more directly from the Klein-Gordon recurrence behind (2.76), avoiding the numerical fit, and test whether log Λ_n − [log c(Δ) + (2Δ−3)log n + n log((1/√2)^4 e^{iπ})] vanishes as 1/n with stable constants. If the fitted form (3.55) survives with controlled subleading corrections, the resummation step is safe; if the constants drift or the power of n changes, recompute the pole exponent in (3.60). Independently, compare the resummed GT(τ) near (3.59) with the boundary limit of a direct numerical solution of the scalar Klein-Gordon equation in Euclidean AdS5-Schwarzschild that imposes regularity at the horizon, to check that the near-boundary coefficients are the physical ones.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Chapter 3's central result is not obtained by an exact resummation: the pole (3.60) follows from (i) computing Λ_n for n≲50 with the near-boundary ansatz (2.76), (ii) fitting the large-n form Λ_n^a = c(Δ) n^{2Δ−3} (1/√2)^{4n} e^{iπ n} (Eq. 3.55), and (iii) approximating the OPE sum by an integral, which produces a logarithmic singularity when τ^4 = (β√2)^4 e^{iπ}. The whole construction therefore inherits the accuracy of that fit. The fit is performed over a finite window, the constant c(Δ) is determined only up to a reported ≈2% over a finite Δ range, and no error bars are given for the exponents in (3.72). Near integer Δ, c(Δ) has poles (3.56), so the finite-Δ regime where the singularity is claimed is precisely where the fit is most delicate. A mild 1/n correction or a slightly different n^{α} would alter the exponent 2Δ−2 in (3.60) and could shift τ_c, invalidating the quantitative match to the bouncing geodesic. This is compounded by the explicitly flagged assumption of Sec. 3.5.3, item 6, that the near-boundary expansion is reliable down to r=0; if that assumption fails, the fitted Λ_n themselves are not the physical OPE data.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This PhD thesis studies thermal two-point functions in holographic CFTs, using the near-boundary expansion as the main tool. Chapter 3 makes the central claim that the stress-tensor sector of a thermal scalar correlator encodes information about the black hole singularity: in d=4, after resumming multi-stress-tensor OPE contributions, the sector develops poles at tau_c = (beta/sqrt{2}) e^{i pi/4} + i k pi/2 with scaling 1/(tau-tau_c)^{2 Delta - 2}, matching bouncing geodesics that reflect off the singularity. Chapters 4-6 compute thermal stress-tensor two-point functions in Einstein and Gauss-Bonnet gravity, extract anomalous dimensions and OPE data for double-stress tensors, identify three universal functions controlling the near-lightcone behavior, and show that saturation of an ANEC makes the corresponding near-lightcone correlator temperature-independent.","tokens_in":61930,"tokens_out":2972,"duration_ms":33778,"significance":"If the Chapter 3 claim is correct, it is a significant step: it provides a concrete boundary observable that encodes the black hole interior singularity, and it links the resummation of the stress-tensor sector to geodesic probes on the second sheet. The result is universal in the sense that it does not depend on the details of the probe operator. The Chapters 4-6 results are also valuable: they provide explicit CFT data for double-stress tensors in holographic theories, verify consistency with conformal collider bounds, and exhibit a clean universality structure near the lightcone, including the temperature-independence that follows from ANEC saturation. The thesis contains many non-trivial checks against geodesic calculations, CFT conformal blocks, and independent OPE results, and the appendices provide substantial technical detail. The central Chapter 3 derivation, however, rests on numerically fitted large-n OPE coefficients without error bars and on an explicitly flagged assumption about the validity of the near-boundary expansion in the interior, so the headline claim is not yet established to the standard expected for a journal publication.","major_comments":[{"comment":"The central singularity claim is obtained from the fitted large-n form Lambda_n^a = c(Delta) n^{2 Delta - 3} (1/sqrt{2})^{4n} e^{i pi n}, with c(Delta) determined only up to a reported ~2% accuracy and with no error bars quoted for the exponents in (3.72). The integral approximation (3.58) then produces the pole (3.60) with exponent 2 Delta - 2. This chain is load-bearing: a subleading 1/n correction or a slightly different power n^{alpha} would change the exponent and could shift the location tau_c, invalidating the quantitative match to the bouncing geodesic result (3.41)-(3.42). The manuscript does not provide an error estimate for the resummation, and the fit window n <= 50 with a crossover at n* = Delta/2 means that the finite-Delta regime where the claim is made is precisely where the fit is most delicate, especially near integer Delta where c(Delta) has poles (3.56). Without an independent check of the large-n asymptotics, or a quantified bound on the 1/n corrections, Eq. (3.60) should be regarded as a numerically motivated conjecture rather than an established result.","section":"Sec. 3.3.3, Eqs. (3.55)-(3.60)"},{"comment":"The author explicitly flags the hidden working assumption that the near-boundary expansion (2.76) reliably solves the Klein-Gordon equation in the region r in (0, infinity). This assumption is load-bearing for the thesis's main claim, because the OPE coefficients Lambda_n are extracted from that expansion; if the near-boundary expansion fails in the interior, the fitted Lambda_n are not the physical OPE data, and the geodesic match in Sec. 3.3.5 is unsupported. The manuscript should supply evidence for this assumption, for example by comparing the near-boundary solution with a full numerical solution of the bulk equation of motion, or by estimating the radius of convergence of the 1/r expansion. The current discussion in Sec. 3.5.3 item 6 only notes the issue for the star example and does not test the Schwarzschild-AdS case where the assumption is actually used.","section":"Sec. 3.5.3, item 6"},{"comment":"The transition from the finite-Delta bouncing singularity to the large-Delta branch point at tau = beta/2 is argued through a noncommutativity of limits, but the argument is not a derivation. In particular, the identification in (3.87), which determines L_[phi phi] and bL from the branch-point function fT(y), is assumed from the geodesic side rather than derived from the OPE data. The large-Delta fit uses Delta = (10^8 + 1)/2 and estimates L_n with 1/Delta corrections, but the quoted uncertainties in (3.72) are small yet unquantified in their effect on the branch-point location and prefactor in (3.75). Given that the agreement with the geodesic result (3.76) is excellent, this is not an objection to the result itself, but the manuscript should state more clearly which parts of Sec. 3.4 are conjectural and which follow from the OPE data alone.","section":"Sec. 3.3.5 and Sec. 3.4.2"}],"minor_comments":[{"comment":"Figure 3.6 plots ratios of explicit Lambda_n to the leading large-n form, but the figure does not show error bars or the n-range used for the fit; adding both would help the reader judge the stability of (3.55).","section":"Sec. 3.3.3, Fig. 3.6"},{"comment":"The fit results log a = 4 log 2 +/- 10^{-6}, b = -7/3 +/- 10^{-4}, log c = -2.050 +/- 10^{-3} are quoted without a statement of the fitting procedure or the covariance of the errors; please specify the fit range and the treatment of 1/n and 1/Delta corrections.","section":"Sec. 3.3.4, Eq. (3.72)"},{"comment":"The relation b_{T_mu nu}/beta^4 = - mu C_T S_4 / 40 introduces mu on the right and mu is also used for the bulk mass parameter; the two uses are related but the notation is confusing. A brief comment distinguishing them would improve readability.","section":"Sec. 2.3, Eq. (2.83)"},{"comment":"The extraction of the anomalous dimensions gamma_J^{(1)} relies on comparing three polarizations with a common solution; please state explicitly whether the solution (4.94)-(4.95) is unique, or whether the remaining freedom in the rho^{(1)} coefficients could change gamma_J^{(1)}.","section":"Sec. 4.3.5, Eqs. (4.94)-(4.95)"}],"recommendation":"major_revision","confidential_remarks":"The thesis is based on four co-authored papers, and the referee report concerns the thesis as submitted. The main unresolved issue is the numerical fitting of the large-n OPE coefficients in Chapter 3, which carries the central claim. The author's own flags in Sec. 3.5.3 item 6 and the absence of error bars on the fit are sufficient grounds for major revision rather than acceptance. The journal should also consider whether the thesis format, with its extensive review material, is appropriate for the journal's scope; the original papers may be the more suitable citable form for the results in Chapters 4-6."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Sam—read this if you care about how black hole singularities show up in boundary correlators. The thesis has a real idea: in a thermal CFT, the stress-tensor sector of a scalar two-point function develops branch-point singularities in complex time exactly where bouncing geodesics in the bulk hit the singularity. The mechanism—resumming multi-stress-tensor exchanges to large order—is new, and if it's right, it settles an old question in holography. But the load-bearing step is a numerical fit, and the thesis doesn't yet control that fit well enough for the match to be called robust.\n\nWhat's genuinely good: the near-boundary expansion is pushed further than before, to spinning TT correlators and to Gauss-Bonnet gravity. The OPE data extraction (anomalous dimensions of double-stress tensors, relations among OPE coefficients) is careful and matches known results where they exist. The large-Δ limit of the stress-tensor sector correctly reproduces the geodesic length to order τ^20, which is a strong check. I also like the GFF example: it shows cleanly why double traces restore KMS and why the order of limits matters. The author is honest about the hidden assumption that the near-boundary expansion is valid down to r=0.\n\nThe soft spot is exactly what the stress-test says. The finite-Δ singularity is built on Λ_n fitted for n up to about 50, with a claimed c(Δ) good to 2%, no error bars on the large-n exponent, and a sum-to-integral approximation. A mild 1/n correction would shift the pole location or the exponent. The fit is most delicate near integer Δ, where c(Δ) has poles, so the claimed singularity lives in the regime where the fit is most fragile. This doesn't kill the idea—the geodesic match in d=4, 6, 8 is suggestive, and the large-Δ limit is solid—but the central result is not yet proven. The thesis also leans on its own earlier papers for some support, which is fine, but it means the reader has to trust the whole chain.\n\nWho should read it: holographers working on thermal correlators or the OPE; they'll find plenty of useful technology and data. The singularity claim needs a serious referee who will demand analytic or independent control of the large-n asymptotics, or at least a rigorous error estimate.\n\nRecommendation: send it to referees. The work is important enough and technically substantial enough to deserve that. But my verdict would be conditional: the singularity claim shouldn't be accepted without the error analysis.","headline":"A technically rich thesis whose headline singularity claim rests on a numerical fit that still needs error control.","tokens_in":62481,"tokens_out":4202,"would_cite":true,"duration_ms":40467,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["11.25.Tq","04.70.Dy"],"model":"deepseek-v4-flash","headline":"The stress-tensor sector of a thermal correlator carries poles that mark the black hole singularity.","keywords":["AdS/CFT correspondence","thermal correlators","black hole singularity","near-boundary expansion","stress-tensor sector","bouncing singularities","averaged null energy condition","Gauss-Bonnet gravity"],"falsifier":"Compute the stress-tensor sector $G_T(\\tau)$ at larger $n$ without relying on the fitted asymptotic form, for example with high-precision recursion or an independent bootstrap calculation, and check whether the pole at $\\tau_c = (\\beta/\\sqrt{2})e^{i\\pi/4}$ survives with the predicted exponent $2\\Delta-2$; alternatively, consider a nonsingular spacetime such as a star that shares the same near-boundary expansion and test whether the same pole appears, since its presence would mean the pole is not a singularity diagnostic.","tokens_in":61289,"feed_emoji":"🕳️","tokens_out":9662,"duration_ms":97606,"temperature":0.7,"pith_summary":"This thesis argues that the deep interior geometry of a black hole, down to its singularity, is encoded in a purely boundary quantity: the stress-tensor sector of a thermal two-point function. Resumming the OPE contributions from the stress tensor and its composites produces branch-point-like poles in complex time, located precisely where bulk geodesics that pass the horizon, bounce off the singularity, and return to the boundary would end. At finite conformal dimension the poles sit at $\\tau_c = (\\beta/\\sqrt{2})e^{i\\pi/4} + ik\\pi/2$ in $d=4$ with exponent $2\\Delta-2$; in the large-dimension limit the same sector reproduces the geodesic proper length and its branch point at $\\tau=\\beta/2$. A second line of results shows that thermal stress-tensor correlators in pure-gravity duals become universal near the lightcone, with the bulk Lagrangian entering only through three parameters, and that saturating the averaged null energy condition makes the correlator independent of temperature. The combined picture makes thermal correlators, especially the stress-tensor sector, sensitive probes of black hole interiors and causality constraints.","feed_headline":"Poles in a thermal correlator expose black hole singularities","feed_subtitle":"Resumming the stress-tensor sector of a CFT reproduces geodesics that bounce off the singularity.","key_machinery":"The load-bearing object is the stress-tensor sector of the boundary OPE: the set of multi-stress-tensor primary operators $[T^n]_J$ (spin $J=0,2,\\ldots,2n$, dimension $dn$ at leading order) and their descendants. On the bulk side the machinery is the near-boundary expansion of the bulk-to-boundary propagator, $\\Phi_T = (r/w^2)^\\Delta\\left(1 + \\sum a_{m,k}^{(n)}\\rho^{2m}w^{2k} r^{-dn}\\right)$, which determines the OPE coefficients $\\Lambda_n$ order by order in $1/r$. The argument is carried by the asymptotic large-$n$ form $\\Lambda_n \\approx c(\\Delta)\\, n^{2\\Delta-3}(1/\\sqrt{2})^{4n}e^{i\\pi n}$, whose resummation by an integral produces a logarithm whose vanishing locus is the pole set; the same OPE data, when the large-$\\Delta$ limit is taken first, yields the branch point of the geodesic length. For the second theme, the machinery is the reduction of linearized Einstein and Gauss-Bonnet perturbations to three channels (scalar, shear, sound), integrated over two spatial directions so the equations collapse to three gauge-invariant PDEs whose near-lightcone solutions are described by three universal functions.","core_discovery":"In a holographic CFT dual to Einstein gravity, the thermal two-point function of a scalar splits into a stress-tensor sector, built from multi-stress-tensor operators $[T^n]_J$, and a double-trace sector. The central claim is that the stress-tensor sector alone knows about the curvature singularity: after resumming the large-$n$ OPE coefficients, the $d=4$ correlator $G_T(\\tau)$ develops singularities at $\\tau_c = (\\beta/\\sqrt{2})e^{i\\pi/4} + ik\\pi/2$ with $G_T(\\tau)\\propto (\\tau-\\tau_c)^{-(2\\Delta-2)}$, precisely the locations and exponents predicted by geodesics that cross the horizon, reflect off the singularity, and return to the boundary. The double-trace sector is then forced to carry the same singularity with the opposite sign, so that the full correlator satisfies KMS and remains analytic in the physical strip; this explains why the bouncing singularity is invisible in the full correlator while still being present in the sector. In the large-$\\Delta$ limit the finite-$\\Delta$ pole is replaced by a branch point of $-(1/\\Delta)\\log G_T(\\tau)$ at $\\tau=\\beta/2$, matching the bulk geodesic length, with the order of the $\\tau\\to\\tau_c$ and $\\Delta\\to\\infty$ limits controlling which description applies. For pure-gravity duals, the thesis further establishes near-lightcone universality of stress-tensor two-point functions in Einstein and Gauss-Bonnet gravity, where three universal functions describe the correlator and the bulk action enters only through corrections to cubic stress-tensor couplings and the thermal one-point function; when an averaged null energy condition is saturated, the near-lightcone correlator takes the vacuum form and becomes temperature-independent.","pith_inferences":["A nonsingular star with the same asymptotic metric shares the same near-boundary data, so the stress-tensor sector should be identical; all information about the interior would then live in the double-trace sector, a prediction that could be tested against an explicit star solution.","The same pole-tracking logic gives an observable signature of singularity resolution: finite-coupling corrections should move or smear the pole at $\\tau_c$, so locating it at higher orders would expose how stringy or loop corrections modify the interior.","The extra singularities noted in the paper for $d=6$ and $d=8$ (at $\\tau=\\beta$ and at $\\frac{\\beta}{2}e^{\\pm i\\pi/8}/\\sin(\\pi/8)$ respectively) may correspond to geodesic families not present in $d=4$, or to non-geodesic saddles; identifying their bulk origin would be a natural extension.","The numerical fit that fixes $c(\\Delta)$ could be replaced by a bootstrap consistency check: truncating the spectrum to multi-stress tensors and double traces and imposing crossing plus KMS should force the predicted pole location and exponent if the claim is correct."],"forward_implications":["A boundary CFT calculation of multi-stress-tensor OPE data can locate the black hole singularity in complex time without reconstructing the bulk geometry.","Double-trace operators are not optional: they are required to restore KMS and analyticity, and their singularity must cancel the stress-tensor pole, so the full correlator's analytic structure determines them from the universal sector.","The map between OPE sectors and bulk geodesics explains the noncommutativity of the large-dimension and near-singularity limits: the finite-$\\Delta$ pole disappears into the branch point at $\\tau=\\beta/2$ as $\\Delta\\to\\infty$.","In pure-gravity duals, the near-lightcone stress-tensor correlator is universal across Einstein and Gauss-Bonnet gravity up to three parameters, so higher-derivative corrections affect it only through those parameters.","Saturating an averaged null energy condition forces the near-lightcone correlator to the vacuum form and makes it temperature-independent, which implies saturation of all higher-spin averaged null energy conditions for that polarization."],"supporting_citations":[{"why":"Introduces bouncing geodesics that reflect off the black hole singularity and the large-$\\Delta$ singular behaviour of the thermal correlator.","marker":"[13]"},{"why":"Provides the geodesic proper-length and momentum-space analysis that the OPE results are matched against.","marker":"[14]"},{"why":"Supplies the near-boundary expansion ansatz that determines the multi-stress-tensor OPE coefficients.","marker":"[30]"},{"why":"Gives the bootstrap computation of the stress-tensor sector OPE coefficients used for cross-checks.","marker":"[56]"},{"why":"Establishes thermalization of multi-stress tensors in heavy states and the thermal block decomposition used in the stress-tensor correlator analysis.","marker":"[46]"},{"why":"Relates stress-tensor contributions in thermal stress-tensor correlators to conformal collider bounds and averaged null energy conditions.","marker":"[114]"},{"why":"Provides the Gauss-Bonnet black hole perturbations, gauge invariants, and central charge used for the higher-derivative analysis.","marker":"[160]"}],"fun_headline_variants":["Stress-tensor sector of thermal CFT sees black hole singularity","Thermal correlator's stress-tensor sector carries singularity imprints","Holographic thermal correlator encodes singularity in stress-tensor sector","Branch points in thermal stress-tensor sector mark black hole singularity"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the near-boundary expansion gives the full stress-tensor sector of the bulk solution from the AdS boundary down to $r=0$, so the boundary OPE data uniquely fixes the analytic structure that corresponds to the interior geometry, and that the numerically fitted large-$n$ OPE coefficients control the resummation that produces the poles.","fun_headline_variants_meta":{"raw":{"variants":["Stress-tensor sector of thermal CFT sees black hole singularity","Thermal correlator's stress-tensor sector carries singularity imprints","Holographic thermal correlator encodes singularity in stress-tensor sector","Branch points in thermal stress-tensor sector mark black hole singularity"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000848,"raw_usage":{"total_tokens":3814,"prompt_tokens":1197,"completion_tokens":2617,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":813,"completion_tokens_details":{"reasoning_tokens":2544}},"tokens_in":813,"tokens_out":2617,"duration_ms":19205,"temperature":1.0,"reasoning_tokens":2544,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T17:07:39.980161+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the stress-tensor sector $G_T(\\tau)$ at larger $n$ without relying on the fitted asymptotic form, for example with high-precision recursion or an independent bootstrap calculation, and check whether the pole at $\\tau_c = (\\beta/\\sqrt{2})e^{i\\pi/4}$ survives with the predicted exponent $2\\Delta-2$; alternatively, consider a nonsingular spacetime such as a star that shares the same near-boundary expansion and test whether the same pole appears, since its presence would mean the pole is not a singularity diagnostic.","supporting_citations":[],"review_version":2}