{"id":"de855c40-c44b-4482-9b5c-7f8b0e0c4616","arxiv_id":"1908.04786","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A screen-source formulation of semi-classical holography yields correlators and bulk reconstruction in flat space and reduces to AdS/CFT at the AdS boundary.","lead":"This paper proposes a way to do holography in general spacetimes by placing the sources that define the dual theory on a screen inside the space, rather than on its boundary. It shows the idea reduces to the standard AdS/CFT recipe at the boundary of AdS and gives concrete formulas for flat space with a spherical screen.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The generality claim rests on an unproved existence/uniqueness premise for the screen-source Green function and homogeneous mode; the paper's own caveats make the central claim conditional.","rationale":"The paper's strongest claim is explicitly a programmatic generalization: the screen-source prescription is supposed to define semi-classical holographic correlators for 'fairly general' spacetimes, reducing to AdS/CFT at the boundary of AdS and giving homogeneous modes for bulk reconstruction outside AdS. For the central formulas (2.3), (3.2), (2.7), and (3.5) to be valid, one needs well-posedness facts about elliptic and hyperbolic PDEs in the Euclidean and Lorentzian sections: existence and uniqueness of the Euclidean Green function with decay at infinity, its analytic continuation to the Feynman propagator, and a complete characterization of regular homogeneous Lorentzian modes. The paper does not prove these facts; it repeatedly uses 'we expect' and 'somewhat imprecisely' (Sec. 1, fn. 1; after Eq. 2.4, fn. 7; Sec. 3, fn. 15), and it explicitly flags the danger of leakage through top/bottom of tube-like regions. These are exactly the places where the prescription could break: if the Euclidean Green function is not unique, the correlators in (2.7)/(3.5) are ambiguous; if leaky modes exist, (3.2) is not the general solution and the state interpretation of phi_h is incomplete. The flat-space R x S^2 example and the AdS limit support the construction but do not cover the claimed generality. This is a conditional issue rather than a demonstrated contradiction, so the reader's CONDITIONAL verdict is appropriate; my concern matches the reader's weakest assumption. The advertised check that extrapolate and differential dictionaries match is largely built into the action choice, which further underlines that the novel generality claim rests on the unproved PDE facts rather than on an independent derivation.","tokens_in":16633,"tokens_out":27205,"duration_ms":311033,"concrete_test":"Compute, by mode sum, the massive scalar Euclidean Green function on Euclidean Schwarzschild with a codimension-one screen at r=R>2M, imposing lim_{r→infty} r G_E = 0; test uniqueness by checking whether a nontrivial regular solution of the homogeneous equation with the same falloff exists. This directly probes the existence/uniqueness premise of (2.2)-(2.4) beyond the flat-space example; a failure would falsify the claimed generality, while success would support but not prove it.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The single most load-bearing concern is the unproved PDE well-posedness premise on which the generality claim rests. To define (2.3) and (3.2) one needs: (i) existence and uniqueness of the Euclidean Green function solving (2.4) with decay at infinity; (ii) its analytic continuation as the Feynman propagator; and (iii) completeness of the homogeneous mode space in (3.2), with no additional leaky modes. The paper does not supply these facts. It states the class of spacetimes 'somewhat imprecisely' (Sec. 1, fn. 1), uses 'we expect' for existence/uniqueness (fn. 7), and flags that solutions in tube-like regions could leak through the top/bottom and change the solution structure (Sec. 3, fn. 15). For a geometry where the Euclidean completion is compact or has multiple ends, or where the Lorentzian tube admits radiation through its top/bottom, the needed Green function need not be unique, its analytic continuation need not be the Feynman propagator, or extra homogeneous modes beyond phi_h can appear. If any one of these fails, formulas (2.7) and (3.5) are not the general screen correlators, and the advertised reduction to AdS/CFT plus the flat-space R x S^2 example do not establish the general prescription. The paper itself concedes the missing proof; as written, the central claim is conditional.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a source-based reformulation of semi-classical holography. Instead of solving a Dirichlet boundary-value problem with boundary data on the holographic screen, the authors solve a bulk wave equation with a source localized on the screen, using the Euclidean Green function that vanishes at infinity. The on-shell action is then written as the integral of the bulk field times the source on the screen, and correlators are obtained by functional differentiation with respect to the source. In Lorentzian signature the authors add a regular homogeneous mode to the sourced solution, interpret it as a state-dependent one-point function, and construct an HKLL-like smearing kernel for bulk reconstruction. They claim the prescription reduces to standard AdS/CFT at the boundary of AdS, works for flat space with an R times S^2 timelike screen, and matches the extrapolate and differential dictionaries. The paper also sketches perturbative interactions and a broader philosophy of describing dynamics via sources on submanifolds.","tokens_in":16910,"tokens_out":6009,"duration_ms":67902,"significance":"If the prescription is correct for a wide class of screens, it provides a concrete semi-classical holographic dictionary beyond AdS, with the notable advantage that the bulk propagator is the standard Feynman propagator and it yields a natural analogue of the normalizable mode. The paper is clearly written and the basic construction is simple and explicit, especially in the flat-space example where the Green functions and mode expansion are given. The main value is as a proposal: it identifies a plausible generalization of the AdS/CFT dictionary and spells out the calculational scheme, including correlators, homogeneous modes, and bulk reconstruction. However, the claimed generality is not established, and some of the advertised checks are definitional rather than substantive.","major_comments":[{"comment":"","section":"Sec. 2, Eq. (2.4), and Sec. 3, Eq. (3.2)"},{"comment":"","section":"Sec. 2, Eqs. (2.5)-(2.7) and Sec. 3, Eqs. (3.3)-(3.5)"},{"comment":"","section":"Sec. 3, text before Eq. (3.2), and Sec. 4.3, Eq. (4.3)"},{"comment":"","section":"Sec. 4.3 and Appendix A"}],"minor_comments":[{"comment":"","section":"Sec. 1, footnote 5"},{"comment":"","section":"Sec. 2, Eq. (2.13)"},{"comment":"","section":"Sec. 4.3, Eq. (4.4)"},{"comment":"","section":"Sec. 5, paragraph after Eq. (5.2)"},{"comment":"","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The paper is a proposal letter whose main claims are considerably broader than what is proved. The reader's and skeptic's concerns about well-posedness are real and are acknowledged in the manuscript itself; the 'extrapolate/differential match' is definitional. That said, the core idea is interesting and the flat-space example is concrete enough to be a useful starting point. The appropriate path is major revision: sharpen the class of spacetimes, fix the homogeneous-mode description, and either provide the flat-space details or clearly label them as preliminary. I would also note that the heavy reliance on 'To appear' companion papers makes the present manuscript difficult to evaluate as a standalone contribution."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Buddy, here's the take.\n\nThe paper is a programmatic proposal for doing semi-classical holography with sources on the screen instead of Dirichlet boundary values. The genuinely new piece is the source prescription: you solve (2.2) with a delta-function source at r=R, use the bulk-to-bulk Euclidean Green function that dies at infinity, and define correlators by varying the on-shell action (2.6). That choice gives you, in Lorentzian signature, a homogeneous mode φ_h that behaves like the AdS normalizable mode, and it opens the door to HKLL-like reconstruction. The explicit flat-space example with an R×S^2 screen is concrete: the two-point function, the mode expansion (4.3), and the kernel (4.4) are all checkable, and the AdS limit correctly reduces to the standard dictionary with the scaling map (2.13). The authors are also honest that the match between extrapolate and differential dictionaries is basically built into the definition of the on-shell action—that's not a hidden circularity, it's just a feature of their setup.\n\nThe soft spot is the one the stress test flags, and the paper itself concedes it. The general prescription requires existence and uniqueness of the Euclidean Green function with decay at infinity, its analytic continuation as the Feynman propagator, and completeness of the homogeneous modes. None of that is proved for general screens. Footnote 7 says 'we expect' and footnote 15 worries about leakage through the top/bottom of a tube-like region. So the central claim is conditional. That's a real gap, but it's an addressable one—it's a matter of PDE analysis, not a contradiction. The appendix's alternate prescription (integrating only to the screen) is less compelling and the resulting correlator (A.5) looks messy, but it's peripheral.\n\nThis is a paper for people working on flat-space holography and bulk reconstruction. It gives a different route from Solodukhin and Li-Takayanagi, and the concrete example makes it worth engaging with. I would not cite it as a proof of anything, but I'd cite it as a formulation.\n\nMy recommendation: send it to peer review. A good referee should push for a sharper statement of the class of spacetimes and at least a plausibility argument for the homogeneous mode, but the paper is honest, the example is real, and the idea deserves to be in the literature. I'd bring it to the reading group.","headline":"Source-on-screen holography is a genuinely different and concretely worked-out prescription, but the advertised generality rests on PDE well-posedness that the authors explicitly leave unproved.","tokens_in":17444,"tokens_out":2947,"would_cite":true,"duration_ms":28414,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"A screen-source prescription for semi-classical holography beyond AdS.","keywords":["semi-classical holography","holographic screen","source prescription","Feynman propagator","homogeneous mode","bulk reconstruction","flat space holography","operator dictionary"],"falsifier":"Take a specific spacetime region with a timelike screen that meets the paper's stated sufficient conditions (e.g., a tube in Schwarzschild exterior whose constant-time intersections are $S^{d-1}$) and solve the Lorentzian scalar wave equation with a screen-localized source. If the regular homogeneous mode does not exist, or if the analytic continuation of the Euclidean Green function is not the Feynman propagator, the prescription's reconstruction and two-point function formulas fail. In the flat-space example, the correlator (3.5) can be computed directly; comparing it with the result of the alternative boundary-value prescription would test whether the two approaches are genuinely inequivalent at finite screen radius.","tokens_in":16429,"feed_emoji":"","tokens_out":8820,"duration_ms":78586,"temperature":0.7,"pith_summary":"This paper proposes a general recipe for semi-classical holography in spacetime regions that are not anti-de Sitter space, with flat space as a primary example. Instead of fixing the boundary values of bulk fields on the screen, the recipe puts localized sources on the screen and solves the bulk wave equation with those sources, using the Euclidean Green function that dies at infinity. The paper argues that this source prescription reproduces standard AdS/CFT when the screen sits at the boundary of AdS, while in more general settings it has two advantages: the analytic continuation of the Euclidean Green function is the Feynman propagator, and a regular homogeneous mode appears that plays the role of the normalizable mode and enables bulk reconstruction. The claims are demonstrated for a scalar field in 3+1 dimensional flat space with a timelike $\\mathbb{R}\\times S^2$ screen.","feed_headline":"Put sources on the screen to extend holography beyond AdS","feed_subtitle":"A new prescription reproduces AdS/CFT at the boundary and gains a homogeneous mode for bulk reconstruction.","key_machinery":"The load-bearing object is the Euclidean bulk-to-bulk Green function $G_E$ that dies at infinity, used as a bulk-to-screen propagator for the source $J_0$, along with the regular homogeneous mode $\\varphi_h$ that appears in Lorentzian signature. The former replaces the Dirichlet Green function, which is defined by vanishing on the screen and whose analytic continuation is not generally the Feynman propagator; the latter is the analogue of the AdS normalizable mode. The smearing kernel for bulk reconstruction is obtained from the spacelike version of this Green function via a Green's theorem identity.","core_discovery":"The central claim is that semi-classical holographic correlators can be computed from a bulk source problem on the holographic screen, rather than a boundary value problem. Concretely, the bulk field satisfies (2.2) with a source $J_0$ localized at $r=R$, the solution is (2.3) built from the bulk-to-bulk Green function $G_E$ that vanishes at Euclidean infinity, and the on-shell action equals $\\int \\sqrt{\\gamma}\\,\\varphi(R,x)J_0(R,x)$. In Lorentzian signature, the Green function analytically continues to the Feynman propagator, and an additional regular homogeneous solution $\\varphi_h$ appears; its on-screen value is the one-point function, and the two-point function is the restriction of the Green function. The extrapolate and differential dictionaries for computing screen correlators coincide, and the prescription reduces to the standard boundary-value AdS/CFT dictionary at the AdS boundary via the scaling map (2.13). Bulk reconstruction proceeds through a smearing kernel derived from the spacelike version of the Green function.","pith_inferences":["If the homogeneous mode exists generically, it could serve as a probe of bulk geometry in settings where the screen is not at infinity, potentially giving a handle on interior physics in black-hole or cosmological spacetimes; the paper only sketches this possibility.","The equivalence of source and boundary-value prescriptions at the AdS boundary suggests there may be a larger class of geometries where a rescaling of the source mimics a boundary value problem; searching for such geometries could sharpen the boundary of where the new prescription is needed.","In flat space, taking the large-$R$ limit of the screen correlators may yield flat-space S-matrix elements in the regime where $E R \\gg 1$, mirroring the AdS/CFT relation to scattering; the paper raises this question but does not establish it."],"forward_implications":["The source prescription gives explicit holographic correlators for flat space with a timelike $\\mathbb{R}\\times S^2$ screen, and perturbative bulk interactions can be added using standard bulk perturbation diagrams.","Bulk reconstruction works in non-AdS regions: the homogeneous mode's boundary value determines the bulk field through a smearing kernel, extending the AdS reconstruction program.","The extrapolate and differential dictionaries coincide in the source prescription, a property that does not hold for the Dirichlet prescription in general.","The prescription reframes mechanics: instead of boundary conditions at a submanifold, one may localize sources on that submanifold and impose fall-off at infinity, a viewpoint with potential applications beyond holography."],"supporting_citations":[{"why":"Defines the AdS/CFT duality that the proposed prescription generalizes.","marker":"[1]"},{"why":"Gives the gauge-theory correlator prescription (differential dictionary) that the screen-source approach reduces to at the AdS boundary.","marker":"[2]"},{"why":"Provides the bulk perturbation formulation and boundary-source dictionary used for comparison.","marker":"[3]"},{"why":"Introduces the normalizable/non-normalizable mode split in Lorentzian AdS that the homogeneous mode mirrors.","marker":"[5]"},{"why":"Supplies the smearing-kernel method for bulk reconstruction, adapted by the paper to screen data.","marker":"[8]"},{"why":"Provides the explicit AdS bulk-to-bulk Green function used to show the reduction to standard AdS/CFT.","marker":"[10]"},{"why":"Defines the spacelike Green function used for the reconstruction kernel.","marker":"[11]"},{"why":"Establishes the extrapolate/differential dictionary match that the paper shows persists in the source prescription.","marker":"[15]"},{"why":"Presents an alternative Dirichlet-based flat-space holography that the paper contrasts with its screen-source prescription.","marker":"[16]"}],"fun_headline_variants":["Screen sources unlock holography beyond AdS","New prescription: sources on screen, not boundary values","Holography from sources, not boundary values","Beyond AdS: sources on screen give correlators","Semi-classical holography via screen-localized sources"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The prescription assumes that for the spacetime regions in question the Euclidean source problem (2.2)-(2.4) has a unique solution with a Green function that dies at infinity, that its analytic continuation is the Feynman propagator, and that a regular homogeneous mode $\\varphi_h$ exists in Lorentzian signature; the paper states these conditions with 'we expect' language and leaves the class of spacetimes imprecise.","fun_headline_variants_meta":{"raw":{"variants":["Screen sources unlock holography beyond AdS","New prescription: sources on screen, not boundary values","Holography from sources, not boundary values","Beyond AdS: sources on screen give correlators","Semi-classical holography via screen-localized sources"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000154,"raw_usage":{"total_tokens":1219,"prompt_tokens":963,"completion_tokens":256,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":579,"completion_tokens_details":{"reasoning_tokens":179}},"tokens_in":579,"tokens_out":256,"duration_ms":7427,"temperature":1.0,"reasoning_tokens":179,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:32:57.569385+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a specific spacetime region with a timelike screen that meets the paper's stated sufficient conditions (e.g., a tube in Schwarzschild exterior whose constant-time intersections are $S^{d-1}$) and solve the Lorentzian scalar wave equation with a screen-localized source. If the regular homogeneous mode does not exist, or if the analytic continuation of the Euclidean Green function is not the Feynman propagator, the prescription's reconstruction and two-point function formulas fail. In the flat-space example, the correlator (3.5) can be computed directly; comparing it with the result of the alternative boundary-value prescription would test whether the two approaches are genuinely inequivalent at finite screen radius.","supporting_citations":[],"review_version":1}