{"id":"4d973925-7e58-4a11-8d49-a0750f571359","arxiv_id":"2411.16542","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A generalized entanglement membrane with an extra bulk-depth degree of freedom correctly captures reflected entropy in 2d CFT, and a relevant deformation restores the ordinary non-degenerate membrane tension.","lead":"This paper derives a generalized entanglement membrane with an extra scalar degree of freedom that correctly captures reflected entropy in 2d CFT, where the standard membrane fails. It also shows how a relevant deformation restores the ordinary membrane and computes information velocities from the membrane picture.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Central claim inherits the unproven EWCS/reflected-entropy duality (3.3); the ℓ/2 vs ℓ distinction is only about reflected entropy if that conjecture holds.","rationale":"The reader's weakest_assumption identifies the EWCS/reflected-entropy duality as the most load-bearing premise. I agree: the paper's key observable distinction (plateau ℓ/2 versus ℓ) is derived from holographic EWCS computations, and the claim that the ordinary membrane fails for reflected entropy is contingent on this holographic dictionary. The paper is careful and internally consistent, with exact BTZ solutions in Appendix C supporting the generalized-membrane results, and the membrane construction is a useful effective description. However, the reflected-entropy interpretation is not self-contained: it relies on a conjecture that, although widely accepted and tested in [30], is not proven here. The concrete CFT replica check of the equilibrium plateau would directly validate or falsify the central claim. Since the reader already assigned CONDITIONAL with HIGH confidence, and my analysis does not identify a new fatal flaw, the verdict remains unchanged.","tokens_in":44626,"tokens_out":23951,"duration_ms":199792,"concrete_test":"Perform the 2d CFT replica computation of S_R(A:B) for the quench setup of Sec. 3.2, in the late-time regime T > (ℓ+D)/2, and check whether S_R/2 saturates to ℓ/2 (as in Eq. (3.12)) or to ℓ (the ordinary-membrane prediction). This directly tests the EWCS conjecture (3.3) in the regime where the paper's new equilibrium EWCS result applies, and would settle whether the generalized membrane is truly necessary for reflected entropy.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central claim that the generalized membrane is needed to capture reflected entropy in 2d CFT rests on the identification S_R(A:B)/2 = EWCS area (Eq. 3.3), cited from [27] but not proven. All comparisons between the generalized membrane (giving plateau ℓ/2) and the ordinary membrane (giving plateau ℓ) are comparisons of EWCS lengths; they become statements about reflected entropy only under this duality. While [30] reported agreement between a 2d CFT replica computation and the BTZ geodesic result for this setup, the present paper's new equilibrium-EWCS result (Sec. 3.4) is the portion most in need of a direct boundary check. If (3.3) failed in the equilibrium regime, the claim that the ordinary membrane fails for reflected entropy would lose its meaning, as the holographic quantity would not be reflected entropy. The footnote 17 caveat about the membrane mapping depending on which EWCS is computed further underscores that the generalized membrane is not yet an autonomous effective theory.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper studies the entanglement membrane description of the time evolution of entanglement in 2d CFTs and their relevant deformations. In 2d CFT the holographic derivation of the membrane yields a degenerate tension function E(v) = 1, and the authors propose a 'generalized membrane' with an extra worldline scalar ξ(u), the logarithm of the bulk depth in BTZ. They show that the generalized membrane reproduces the exact BTZ geodesic results for displaced half-spaces (App. B) and that, while entanglement entropy is insensitive to the extra degree of freedom, the reflected entropy (computed holographically via the entanglement wedge cross section, Eq. (3.3)) is not: the generalized membrane yields a plateau SR/2 = ℓ/2 (Eq. (3.12)), whereas the ordinary membrane gives ℓ (Sec. 4.3). The equilibrium EWCS underlying the plateau is constructed in Sec. 3.4 and checked against the exact solution in App. C.2. The paper further shows that a relevant deformation (hairy BTZ with Kasner interior) gaps out ξ and produces an ordinary non-degenerate tension, computed numerically and in conformal perturbation theory (Sec. 5), proposes a membrane description of reflected entropy in d > 2 (Eq. (4.12)) compared with numerical EWCS, and provides a membrane derivation of the information velocity of [1] (Sec. 6).","tokens_in":44817,"tokens_out":54385,"duration_ms":466185,"significance":"The central claim — that the ordinary degenerate membrane fails for reflected entropy in 2d CFT and that an extra worldline degree of freedom is required — is sharp and, if upheld, is a significant step in understanding the limits of the membrane effective theory. The paper is technically strong: the scaling-limit membrane results are checked against exact geodesic solutions in Apps. B and C; the small-deformation perturbation theory (Sec. 5.2) is analytic and internally consistent (notably, the λ² log λ terms cancel in the constraint E(vB) = vB); the construction has no free parameters; and the CFT twist computation in App. B.6.2 independently verifies the entropy result. The ℓ/2 vs ℓ plateau distinction (Sec. 4.3) is a concrete, falsifiable prediction. The main caveats are that the reflected-entropy interpretation inherits the EWCS conjecture (3.3), the new equilibrium plateau lacks a direct boundary computation, the generalized membrane is acknowledged (footnote 17, Sec. 7) not to be an autonomous effective theory, and the d > 2 numerical support (Fig. 14, App. D) is unquantified.","major_comments":[{"comment":"Eq. (3.12) is internally inconsistent as printed. The two EWCS candidates have lengths T − D/2 (Eq. (3.5); exact version (C.8)) and ℓ/2 (Eq. (3.11); exact version (C.22)), so the minimal surface switches at T = (ℓ + D)/2, where the two lengths are equal. As printed, (3.12) keeps the linear branch until T = ℓ + D/2, where that branch has the value ℓ, and then asserts a plateau ℓ/2 for T > ℓ + D/2, producing a downward discontinuity of size ℓ/2. The correct plateau onset is (ℓ + D)/2, which is also what follows from the condition µB ≥ 0 in (C.23) and from taking the vE = vB = 1 limit of Eq. (4.12) with the plateau value ℓ/2. Please correct (3.12), the Fig. 11 caption, and any downstream statements that use this transition time.","section":"Sec. 3.4, Eq. (3.12)"},{"comment":"The headline claim — that the ordinary membrane fails for reflected entropy in 2d CFT and that the generalized membrane is required — is a statement about reflected entropy only if the EWCS/reflected-entropy duality (3.3) holds. That duality is cited to [27] and is a conjecture. The only boundary CFT computation in the paper (App. B.6.2) verifies the displaced-half-space von Neumann entropy, not the EWCS; the non-equilibrium EWCS is covered by the CFT computation of [30], but the new equilibrium plateau of Sec. 3.4 is not. I recommend either providing (or citing) a CFT replica computation for the equilibrium regime — feasible along the lines of App. B.6.2 — or, at a minimum, stating explicitly in the abstract and in Sec. 4.3 that the ℓ/2 plateau is a prediction of the holographic EWCS under conjecture (3.3).","section":"Sec. 3.1, Sec. 3.4, App. B.6.2"},{"comment":"The paper's own caveats are in tension with the unqualified abstract claim that 'in order to correctly capture the reflected entropy in 2d CFT, one needs to add an additional degree of freedom.' Footnote 17 concedes that reproducing the heuristic picture of [30] for the reflected setup requires a bulk-to-boundary mapping that depends on which EWCS is being computed, and Sec. 7 concedes that no direct map of ξ(u) to CFT quantities is known and that no random-circuit interpretation exists. Moreover, the generalized membrane is derived from, and tested against, the same BTZ geodesic action (compare Eq. (2.9) with Apps. B and C), so within this paper it is a scaling-limit reorganization of the holographic computation rather than an independent effective theory. I recommend rescoping the abstract and introduction accordingly, and identifying a concrete test of autonomy (e.g., the footnote-17 setup or a dual-unitary circuit analog).","section":"Sec. 3.4 footnote 17, Sec. 7"},{"comment":"The numerical support for the d > 2 membrane prediction (4.12) is not quantified. No error bars are given; the data in Fig. 14 are for O(1) values of D (D ≤ 3.3), far from the D ≫ 1 scaling limit; and App. D reports that 'precise numerical agreement remains elusive' while Sec. 4.2 states the membrane description is 'surprisingly accurate.' The apparent systematic discrepancies in the plateau and in the onset time (Fig. 39) should be reconciled with the claimed agreement: either provide error bars, a convergence study in 1/D, or an explicit statement of which features of (4.12) are validated. Because the comparison in Sec. 4.3 (Fig. 15) uses (4.12) for the d > 2 curve, the strength of the ℓ vs ℓ/2 distinction partly rests on this unquantified numerical check.","section":"Sec. 4.2, App. D, Figs. 14 and 39"}],"minor_comments":[{"comment":"The jump size is written as ΔSR/(2D) = (1/2)(1 − vE/vB)D, which is dimensionally inconsistent; it should be ΔSR/(2D) = (1/2)(1 − vE/vB) (equivalently, ΔSR/2 = (D/2)(1 − vE/vB)).","section":"Eq. (4.13)"},{"comment":"The text gives the minimum of the potential as ϕ* = 3√2/µ, but the Fig. 20 caption quotes ϕ* = 2.12 for µ = 1, which corresponds to ϕ* = 3/√(2µ); the formula in the text appears to have a factor of 2 error.","section":"Sec. 5.3"},{"comment":"The statement that Im t = −iπ in the thermal double is attributed to reference [38] (Klebanov and Witten); the appropriate reference for the Schwarzschild-time continuation is [21] (Hartman and Maldacena) or a related paper on BTZ geodesics.","section":"App. B.4.2 (citation)"},{"comment":"The rescaling 'introduce p ≡ e^{−2ξ}p' reuses the symbol p for the rescaled momentum; a distinct symbol (e.g., p̃) would avoid confusion, and the subscript on ξp should be typeset explicitly.","section":"Eq. (2.11)"},{"comment":"The Fig. 39 caption contains the typo 'Comparisson' (should be 'Comparison'); several figure captions also have a stray space before the colon in 'Figure X :'.","section":"App. D, Fig. 39"},{"comment":"The branch selection for the parametric curves in Eq. (6.12) ('multiple solutions for λ') is discussed only in one sentence; a brief explanation of how the physical branch is identified would improve reproducibility.","section":"Sec. 6.2"},{"comment":"A brief statement of numerical method and code/data availability would strengthen reproducibility of the EWCS numerics.","section":"Sec. 4.1, App. D"}],"recommendation":"major_revision","confidential_remarks":"This is a careful and technically accomplished holography paper, and I would not want the referee report to read as hostile: the exact solutions in Apps. B and C and the analytic RG-flow analysis in Sec. 5 are strengths. The forcing issue for the verdict is the gap between the abstract's assertive framing and the manuscript's own acknowledged limitations (conjecture (3.3), footnote 17, Sec. 7), together with the concrete inconsistency in Eq. (3.12) that I flag in Major Comment 1. The latter is clearly fixable. On the numerics: I would ask the editor to encourage the authors to include error bars or a convergence statement, since the d > 2 part of the paper currently rests on unquantified numerics. The paper's fit with JHEP is good, and I expect it to be citable; with the requested rescoping and corrections it should be publishable."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This paper is worth a serious referee. The genuinely new piece is the generalized membrane: in 2d CFT you keep the bulk-depth coordinate ξ as a dynamical worldline scalar, and this extra degree of freedom changes the reflected entropy plateau from ℓ/2 to ℓ when compared to the ordinary degenerate-tension membrane. That is a concrete, falsifiable-in-principle distinction, and it is not in the earlier membrane literature. The paper also gives a systematic interpolation to the ordinary non-degenerate membrane through a relevant deformation, with analytic small-coupling expansions plus numerics for the hairy BTZ/Kasner geometry.\n\nWhat earns credit: the scaling-limit derivations in Secs. 2 and 3 are backed by exact solutions in Appendices B and C, and the half-space entropy is cross-checked against a direct CFT twist calculation (B.6.2). That is real evidence, not just a matching exercise. The derivation of the membrane tension in the deformed CFT uses the standard holographic machinery carefully, including near-horizon and deep-interior expansions that satisfy the known consistency conditions on E(v). The citation pattern also looks right: the prior membrane work [4] and the heuristic reflected-entropy picture [30] are properly credited.\n\nThe main caveat is exactly the one the stress-test note flags: the reflected-entropy interpretation of the ℓ/2 plateau depends on Eq. (3.3), the EWCS/reflected-entropy duality from [27]. The paper does not prove that conjecture, and if it failed the plateau would be a statement about entanglement wedge cross sections, not about reflected entropy. I don't treat this as a fatal flaw—it is the standard working hypothesis in holographic entanglement, and the paper is explicit about using it—but a referee should ask for a crisp statement that the boundary reflected entropy prediction is conditional on (3.3), and ideally for a direct CFT replica check of the plateau. The other soft spots are smaller: the d>2 numerical comparison (Fig. 14) has no error bars and no shipped code, and the authors themselves note the small-D limitations; footnote 17 admits the membrane-to-bulk mapping depends on which EWCS one computes, so the generalized membrane is not yet an autonomous effective theory; and the field-theoretic meaning of ξ remains open, as the authors say in the outlook. None of these undercut the core holographic derivation.\n\nMy take: send it to peer review. The exact geodesic and CFT checks are enough that this deserves referee time even if the final version should sharpen the conjecture-dependence and, if possible, add a boundary calculation of the plateau. I would also bring it to reading group—the ξ picture is likely to be influential.","headline":"A genuinely new generalized membrane for 2d CFT with a sharp reflected-entropy prediction; the central claim inherits the standard EWCS conjecture, but the exact-checked core deserves refereeing.","tokens_in":45368,"tokens_out":2592,"would_cite":true,"duration_ms":27590,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["81T40","83C57","81P40"],"pacs":["04.70.-s","11.25.Tq","03.67.Mn"],"model":"deepseek-v4-flash","headline":"For two-dimensional CFTs, the entanglement membrane must carry an extra worldline degree of freedom to reproduce reflected entropy, which the ordinary degenerate-tension membrane gets wrong by a factor of two at late times.","keywords":["entanglement membrane","reflected entropy","two-dimensional CFT","entanglement wedge cross section","RG flow","BTZ black hole","Kasner universe","information velocity"],"falsifier":"Compute the reflected entropy after a global quench in a 2d CFT directly, without invoking the entanglement-wedge-cross-section conjecture, for a bipartition made of intervals of length $\\ell$ separated by $D\\ll\\ell$, and compare the late-time plateau of $S_R/2$ with the predicted $\\ell/2$; a plateau at $\\ell$ would falsify the generalized-membrane prediction. On the holographic side, one can also check the small-$\\phi$ expansion of the membrane tension, e.g. $v_B = 1 - \\frac{3\\pi(3\\pi-8)\\Gamma(3/4)^2}{2\\Gamma(1/4)^2}\\bar\\kappa^2$, against a direct conformal-perturbation-theory computation.","tokens_in":121,"feed_emoji":"🔗","tokens_out":7251,"duration_ms":125799,"temperature":0.7,"pith_summary":"This paper argues that the standard entanglement membrane—a minimal-curve description of how entanglement spreads in chaotic systems—cannot, on its own, account for reflected entropy in two-dimensional conformal field theories. The authors show holographically that a correct membrane description of a 2d CFT must carry an extra scalar degree of freedom on its worldline, the bulk depth $\\xi=\\log z$, in addition to the membrane shape. With this extra degree of freedom, the membrane reproduces the entanglement-wedge-cross-section result for reflected entropy after a quench, including a late-time plateau $S_R/2=\\ell/2$ that the ordinary degenerate-tension membrane would instead predict as $\\ell$. They further show that a relevant deformation of the CFT, dual to a three-dimensional black hole with a Kasner interior, gaps out the extra degree of freedom and recovers the ordinary non-degenerate membrane tension $E(v)$. If this is right, the distance between the generalized and ordinary membrane shows up not in entanglement entropy itself but in finer correlation measures, and 2d CFTs sit at a special point in the space of chaotic systems.","feed_headline":"2d CFTs need an extra membrane field to get reflected entropy right","feed_subtitle":"The plateau value S_R/2 = ℓ/2 distinguishes the generalized from the ordinary membrane in 2d CFTs.","key_machinery":"The load-bearing object is the generalized membrane Lagrangian $L_{\\rm eff} = \\frac12 e^{-2\\xi}(\\dot x^2 - 1 - \\dot\\xi^2)$ obtained from the extremal-surface area functional in a three-dimensional black hole at large depth $z=e^\\xi$, with degenerate tension $E(v)=1$ to leading order. The extra field $\\xi(u)=\\log z(u)$ is a worldline degree of freedom that encodes the bulk radial position of the extremal surface; it must be kept because in $d=2$ no finite plateau in $z$ exists and the extremal surface reaches exponentially deep into the black hole. The paper's main computations map the entanglement-wedge cross section to piecewise generalized-membrane configurations—non-equilibrium vertical segments and equilibrium plateau-plus-null segments—and evaluate their lengths. For the interpolation, a relevant deformation is modeled by an Einstein-scalar action whose interior solution is a Kasner universe; small-$\\phi$ perturbation theory yields analytic expansions for $v_B$, $v_E$, and $E(v)$, and adding a $\\phi^4$ interaction provides a UV-to-IR RG flow with degenerate tensions at both ends.","core_discovery":"On the paper's own terms: in a 2d CFT the entanglement membrane must be promoted to a generalized membrane carrying an extra worldline field $\\xi(u)$ that records how deep the bulk extremal surface goes behind the horizon. When reflected entropy is computed holographically as half the entanglement wedge cross section, the generalized membrane gives $S_R/2 = \\max\\{0,\\, T-D/2\\}$ during linear growth and saturates at $S_R/2 = \\ell/2$. The ordinary membrane with degenerate tension $E(v)=1$ incorrectly predicts the same linear-growth slope but a plateau $\\ell$; the plateau mismatch survives even in the limit where the two theories agree on entanglement entropy. The paper also claims that adding a relevant scalar deformation, dual to a planar three-dimensional black hole with scalar hair and an interior Kasner universe, makes the $\\xi$ degree of freedom gapped out, and the generalized membrane continuously reduces to the ordinary membrane with a non-degenerate $E(v)$. In higher dimensions the ordinary membrane does capture reflected entropy, so the extra degree of freedom is special to $d=2$.","pith_inferences":["Our inference: if the generalized membrane's extra field really encodes the infinite-dimensional symmetry charges of the 2d CFT—an interpretation the paper speculates on but does not prove—then any chaotic 1+1d system with degenerate tension should show the same $\\ell/2$ plateau, a prediction testable in dual-unitary circuits.","Our inference: the paper leaves the random-circuit counterpart of the generalized membrane open; a concrete extension would be to compute reflected entropy in a dual-unitary circuit and look for the worldline degree of freedom as a delocalized domain wall.","Our inference: the RG interpolation suggests a universal crossover in which a weakly deformed 2d CFT behaves like a 2d CFT at early times and like a higher-dimensional chaotic system at late times, with the crossover controlled by $1/\\xi_*$.","Our inference: since the bulk-to-boundary projection depends on which entanglement wedge cross section is being computed, a fully autonomous effective theory would need a selection rule for the projection; until then the generalized membrane is best read as a reorganization of the holographic computation."],"forward_implications":["Reflected entropy, not entanglement entropy, is the observable that distinguishes the generalized 2d membrane from the ordinary one: the plateau $S_R/2$ is $\\ell/2$ rather than $\\ell$.","Adding a relevant deformation to a 2d CFT interpolates continuously from generalized to ordinary membrane theory, with the extra degree of freedom gapped out at a scale set by $e^{-2\\xi_*} \\sim \\lambda^2/(2\\pi^2)$.","In $d>2$ chaotic systems, reflected entropy after a quench is correctly captured by the ordinary membrane, with a plateau $\\ell$ and a jump at the transition that vanishes only when $v_E=v_B$.","The information velocity $v_I(f)$ follows from the Legendre transform of the membrane tension; for spheres with chemical potential it saturates to $v_B$ above a critical $f_c$.","Saturated extremal surfaces project to butterfly-velocity cones of slope $v_B$ in the membrane picture, fixing earlier imprecise treatments of near-horizon membrane sections."],"supporting_citations":[{"why":"Supplies the holographic conjecture that reflected entropy equals the entanglement wedge cross section, which is the load-bearing input for all reflected-entropy results.","marker":"[27]"},{"why":"Derives the entanglement membrane from holography and fixes the tension-function constraints that the generalized membrane extends.","marker":"[4]"},{"why":"Provides the three-dimensional black hole geodesic technology and the end-of-the-world brane setup used for the quench computations.","marker":"[21]"},{"why":"Gives the earlier 2d CFT and heuristic membrane computation of reflected entropy that the generalized membrane reproduces.","marker":"[30]"},{"why":"Provides the conformal-perturbation-theory model of relevant deformations whose hydrodynamic emergence motivates the RG-flow comparison.","marker":"[13]"},{"why":"Defines information velocity and supplies the numerical results that the membrane computation reproduces.","marker":"[1]"}],"fun_headline_variants":["Reflected entropy in 2d CFT needs a generalized membrane","Extra membrane field required for reflected entropy in 2d CFT","Generalized membrane fixes reflected entropy in 2d CFT","Membrane for reflected entropy needs extra field in 2d CFT","2d CFT reflected entropy: extra membrane degree of freedom"],"cache_read_input_tokens":47488,"weakest_assumption_plain":"The argument rests on the holographic conjecture that reflected entropy equals the entanglement wedge cross section (from [27]), plus the assumption that the bulk-to-boundary membrane map is well-defined even though the paper concedes in footnote 17 that the map depends on which entanglement wedge cross section is computed.","fun_headline_variants_meta":{"raw":{"variants":["Reflected entropy in 2d CFT needs a generalized membrane","Extra membrane field required for reflected entropy in 2d CFT","Generalized membrane fixes reflected entropy in 2d CFT","Membrane for reflected entropy needs extra field in 2d CFT","2d CFT reflected entropy: extra membrane degree of freedom"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000423,"raw_usage":{"total_tokens":2164,"prompt_tokens":933,"completion_tokens":1231,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":549,"completion_tokens_details":{"reasoning_tokens":1141}},"tokens_in":549,"tokens_out":1231,"duration_ms":8094,"temperature":1.0,"reasoning_tokens":1141,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T12:59:13.480049+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the reflected entropy after a global quench in a 2d CFT directly, without invoking the entanglement-wedge-cross-section conjecture, for a bipartition made of intervals of length $\\ell$ separated by $D\\ll\\ell$, and compare the late-time plateau of $S_R/2$ with the predicted $\\ell/2$; a plateau at $\\ell$ would falsify the generalized-membrane prediction. On the holographic side, one can also check the small-$\\phi$ expansion of the membrane tension, e.g. $v_B = 1 - \\frac{3\\pi(3\\pi-8)\\Gamma(3/4)^2}{2\\Gamma(1/4)^2}\\bar\\kappa^2$, against a direct conformal-perturbation-theory computation.","supporting_citations":[],"review_version":1}