{"id":"0459bd92-8e7d-422f-ac6b-0b1f679cca3a","arxiv_id":"1908.03351","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For double local operator excitations in pure 2D CFTs, the late-time entanglement entropy equals the sum of two single-quench results plus a negative c/6 log((l_B - l_A)/(t - l_A)) interaction term.","lead":"This paper studies how entanglement entropy grows when two local operators are excited in a two-dimensional conformal field theory. It finds an extra, negative contribution beyond the sum of the two separate excitations, which the authors interpret as a signature of gravitational attraction in the dual AdS spacetime.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The interaction term (25) rests entirely on the unproved n→1 limit of a six-point conformal block, Eq. (23), with z1 and z2 undefined; this must be independently verified before the result is accepted.","rationale":"The reader's weakest_assumption matches my analysis: Eq. (23), with the supporting factorization (22), is the load-bearing, unproved input. I have considered other possible weaknesses. The 'pure CFT' assumption is clearly stated and is the same condition under which the single-excitation results were derived; it is not new. The gravitational interpretation in Section IV is explicitly heuristic and does not feed back into the derivation of (25). The central-charge independence of the interaction term is an output, not an input. The paper is honest about the limitation that the remaining block 'can not be evaluated for general n' and that only the n→1 limit is used. However, no derivation of that limit is supplied, and the undefined z1,z2 prevent an independent check of the boundary-to-cross-ratio mapping. This is enough to make the result conditional: the formula is plausible and likely correct given the companion paper's framework, but it is not established by the present text. The verdict CONDITIONAL is appropriate; my read does not change it.","tokens_in":8107,"tokens_out":10175,"duration_ms":111347,"concrete_test":"Independently compute the six-point vacuum conformal block in the Regge limit for Liouville CFT (a pure CFT with c>1) using the fusion matrix / DOZZ kernel, explicitly defining z1 = 2iϵ/(t−l_A) and z2 = 2iϵ/(t−l_B). Verify that the n→1 limit of the double-residue term is exactly [z1(z2−z1)]^{−2hσn} up to a prefactor that equals 1 at n=1. In particular, check that there is no extra prefactor f(n) with f(1)=1 but f′(1)≠0, which would shift the interaction term by a constant, and no power dependence different from (23), which would change the coefficient of log((l_B−l_A)/(t−l_A)) in (25).","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central result (25) is obtained in three lines (Eqs. (22)-(24)) from a six-point correlator in the Regge limit. The decisive step is (23): the leftover conformal block after taking two residues at α=2α_n is asserted to equal [z1(z2−z1)]^{−2hσn} in the n→1 limit, with no proof, and the cross ratios z1 and z2 are never defined in the paper. The factorization (22) that produces this block is also imported without proof from arXiv:1905.02191. The entire interaction term −(c/6)log((t−l_A)/(l_B−l_A)) in (25) comes from the exponent −2hσn in (23): after multiplying by 1/(1−n) and taking n→1, hσn≈(c/12)(n−1) yields the coefficient c/6. If (23) is missing a prefactor f(n) with f(1)=1, f′(1)≠0, or if the power of z1(z2−z1) is not exactly −2hσn, the interaction term changes. Because this is the only source of the new physics claimed, the validity of (25) is entirely contingent on (23).","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the time evolution of entanglement entropy after a double local operator excitation in two-dimensional conformal field theories with central charge c > 1, restricted to 'pure' CFTs (no chiral primaries). The main claim is Eq. (25): in the late-time Regge limit t > l_B, the excitation contribution to the half-space entanglement entropy is not the sum of two independent single-excitation contributions but acquires an additional negative term, (c/6) log((l_B - l_A)/(t - l_A)). This term is independent of the operator weights and is exact in the central charge. The derivation is based on a six-point twist-field correlator, evaluated by taking repeated Regge-limit residues of the Virasoro conformal block, using the fusion/monodromy technology introduced in the companion paper arXiv:1905.02191. The paper further proposes a gravitational interpretation: the negative interaction term reflects the attractive force between two particles in AdS.","tokens_in":8316,"tokens_out":3405,"duration_ms":37087,"significance":"If the central result is correct, it is a valuable and surprising universal statement: entanglement entropy is claimed to detect interactions between local excitations in holographic CFTs in a way that is nonperturbative in the central charge and independent of the operator dimensions. This goes beyond the existing single-quench results and beyond the sum rule known for rational CFTs, and it provides a sharp, falsifiable prediction (a negative logarithmic correction) that could be checked by bootstrap or holographic methods. The paper also makes a concrete connection to gravitational attraction and to the classical black-hole merger entropy growth. The main caveat is that the decisive technical step, Eq. (23), is not derived in the paper but only imported from the authors' companion work; the strength of the claim therefore depends entirely on the reliability of that external result.","major_comments":[{"comment":"The central step of the derivation is the assertion that, after the two Regge-pole residues and in the limit n -> 1, the remaining six-point conformal block reduces to [z1 (z2 - z1)]^{-2 h_sigma_n}. This is the only source of the interaction term in Eq. (25), because the exponent -2 h_sigma_n, together with h_sigma_n ~ (c/12)(n-1), produces the coefficient c/6 after the 1/(1-n) prefactor. The manuscript gives no proof or derivation of this limit, and the cross ratios z1 and z2 are never defined. A prefactor f(n) with f(1)=1 but f'(1) != 0, or a power that differs from -2 h_sigma_n by any fixed shift, would change the interaction term. The authors should either provide a self-contained derivation (an appendix would be appropriate) or give an exact reference to the companion paper arXiv:1905.02191 with precise equation and theorem numbers, including the definitions of z1 and z2 and the channel in which the block is evaluated.","section":"Section III"},{"comment":"The factorization of the six-point correlator into two Regge residues times a leftover conformal block is assumed without proof. This is not a trivial step: it requires the monodromy transformation of a six-point Virasoro block in the (Vir)^n / Z_n orbifold theory, and it must hold for all values of the central charge if the claim of c-exactness in Eq. (25) is to be substantiated. The paper states that this result follows from the methods of [1], but it does not specify which statement in [1] applies, what its regime of validity is, or whether subleading terms in the monodromy integral have been dropped. Since the interaction term is the whole new physics claimed, this factorization is load-bearing and needs a clear justification or an explicit pointer to a proof in the companion paper.","section":"Section III"},{"comment":"The order of limits in the derivation is not made precise. The text says 'in the late time limit (i.e., the Regge limit)' and then 'epsilon -> 0 and n-1 -> 0', but it does not specify whether one first takes epsilon -> 0 and then n -> 1, or whether the limits commute. The exponent of (2i epsilon) in Eq. (24) and the appearance of the factor (l_B - l_A)/(t - l_A) could in principle receive corrections if the small-epsilon and small-(n-1) limits interact. The authors should state the exact order of limits and argue that higher-order terms vanish uniformly in n, especially because the final result is claimed to be exact in c and holds for t > l_B without other approximations.","section":"Section III"}],"minor_comments":[{"comment":"The symbols z1 and z2 are introduced in Eq. (23) without definition. Even if the companion paper is referenced, the present manuscript should at least define these cross ratios and indicate how they are related to the physical positions l_A, l_B, t, and epsilon.","section":"Section III"},{"comment":"The definition of the double-excited state in Eq. (5) contains an apparent redundancy: the factor e^{epsilon H + i H t} e^{-epsilon H - i H t} equals the identity, so the operator O_A is inserted at the same time as O_B. This may be a typographical artifact, but it makes the state ambiguous; please rewrite the expression so that the intended time evolution of both operators is clear.","section":"Section I"},{"comment":"The generalizations to a finite interval and to a circle are stated without derivation. Since they are plausible and can likely be obtained from the methods of [26], it would be helpful to outline the conformal-transformation argument that leads from Eq. (25) to Eqs. (26) and (27), even in a footnote.","section":"Section III"},{"comment":"The phrase 'pure CFTs' is central to the validity of the result but is only defined in the introduction as CFTs with c > 1 and without chiral primaries. Since the conclusion that the dominant residue comes from alpha = 2 alpha_n is nontrivial, a brief explanation of why the vacuum contribution is absent in this class would improve readability.","section":"General"}],"recommendation":"major_revision","confidential_remarks":"The paper is a short letter that relies heavily on the companion paper arXiv:1905.02191, which is not independently verifiable from the present manuscript. The decisive Eq. (23) is not proved here, and the cross ratios are undefined. I recommend major revision rather than rejection because the gap appears fixable: the authors can supply a derivation or a precise citation to the companion work. I also suggest that the editor consider whether this manuscript can stand alone without relying on an unreviewed companion, or whether the two papers should be reviewed together."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nThe paper's main result is a concrete prediction: for a double local operator excitation in a pure CFT (c>1, no chiral primaries), the entanglement entropy is the sum of the two single-excitation contributions plus (c/6) log((l_B-l_A)/(t-l_A)) in the late-time limit. That negative interaction term is new, exact in central charge, and independent of the operator dimensions. If it holds, it gives a clean separation between holographic CFTs and RCFTs, where a sum rule applies. The sign matches the intuitive picture that gravitational attraction slows entanglement growth, and the authors are explicit that this is their main result.\n\nWhat the paper does well: the logic is straightforward, the stated restrictions (pure CFTs, vacuum exchange) are clear, and the generalization to an interval and to a circle (Eqs. 26, 27) are useful. The discussion of a possible gravity dual is honest—they note that a two-particle merger reproduces the late-time log t but not the constant terms, so the bulk interpretation is left open.\n\nThe soft spot is exactly where the stress-test lands. The decisive step is Eq. (23): after two Regge residues, the remaining six-point block is asserted to equal [z1(z2-z1)]^{-2h_{sigma_n}} in the n->1 limit, with no derivation and with z1,z2 never defined. Eq. (22) is imported from the authors' companion paper and not independently verified here. Since the coefficient c/6 in (25) comes from that exponent, the main formula rests on this unproved limit. I would not call it a fitted or assumed result—the derivation is a boundary computation, not an input—but it is a genuine gap in a short letter.\n\nIf I were refereeing, I would ask for a proof of (23) or an explicit pointer to the derivation in 1905.02191, plus definitions of z1,z2. A prefactor in (23) that goes to 1 at n=1 with nonvanishing derivative would only shift constant terms, so the qualitative log-time behavior is robust; the dangerous error is a different exponent, which would change the coefficient. I lean towards the formula being right, but I would not rely on it without checking that block limit.\n\nThis paper is for people working on local quenches, entanglement, and bulk-boundary signatures. It is a good reading group candidate because the gap is concrete and checkable. I would send it to peer review: a specialist can settle whether (23) is correct, and if it is, this is a publishable result.","headline":"New exact-in-c interaction term for double local quenches; the physics is suggestive, but Eq. (23) needs a proof before the coefficient is trusted.","tokens_in":8916,"tokens_out":4899,"would_cite":true,"duration_ms":51221,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["11.25.Hf","03.67.Mn"],"model":"deepseek-v4-flash","headline":"A double local quench in a pure 2D CFT produces a negative interaction term in the entanglement entropy, so two excitations do not simply add.","keywords":["entanglement entropy","local quench","double excitation","conformal field theory","Regge limit","conformal block","monodromy matrix","interaction measure"],"falsifier":"Numerically evaluate the six-point function in Eq. (21) for a specific pure CFT at finite replica number $n$ and take the $n\\to1$ limit using available conformal-block numerics; if the leftover factor is not $[z_1(z_2-z_1)]^{-2h_{\\sigma_n}}$, the predicted interaction term $\\frac{c}{6}\\log\\frac{l_B-l_A}{t-l_A}$ would be modified. A simpler check is to measure the late-time slope of $\\Delta S_A[O_A;O_B]$: Eq. (25) predicts a universal $\\frac{c}{6}\\log t$ growth, whereas the naive sum rule predicts twice that slope.","tokens_in":7826,"feed_emoji":"🔗","tokens_out":13456,"duration_ms":124018,"temperature":0.7,"pith_summary":"The paper asks whether entanglement entropy can measure the interaction between two quantum excitations, not just their individual propagation. In a two-dimensional conformal field theory with central charge $c>1$ and no extra conserved currents (a 'pure' CFT), it computes the late-time entanglement entropy after two local operators act on the vacuum at positions $l_A<l_B$. The result is Eq. (25): the entropy is the sum of the two single-excitation contributions plus an extra term $\\frac{c}{6}\\log\\frac{l_B-l_A}{t-l_A}$, which is negative for $t>l_B$. Because the correction is exact in $c$ and independent of the operators' conformal weights, the authors read it as the entanglement-side signature of the attractive gravitational force between the two particles in the holographic dual. This makes entanglement entropy a direct probe of interactions in strongly coupled systems.","feed_headline":"Double quench in 2D CFTs shrinks entanglement below the sum rule","feed_subtitle":"The extra term is exact in central charge and signals the attractive pull of gravity.","key_machinery":"The machinery is the monodromy/fusion approach to multi-point conformal blocks in the cyclic orbifold CFT $\\mathcal{M}^n/\\mathbb{Z}_n$. The six-point correlator of Eq. (21) is decomposed into conformal blocks; the Regge limit picks out the twist-operator exchange and, through the monodromy matrix $M^{(n)}$, converts the correlator into a product of two single-excitation residues (one for $O_A$, one for $O_B$) times a leftover six-point block. The crucial step is Eq. (23): in the $n\\to1$ limit this leftover block is asserted to factor as $[z_1(z_2-z_1)]^{-2h_{\\sigma_n}}$, which turns distances into the logarithms of Eq. (25). The same approach also yields the finite-interval and circle generalizations in Eqs. (26) and (27).","core_discovery":"On the paper's own terms, the central discovery is that the sum rule for entanglement entropy after multiple local excitations fails in pure CFTs, and the failure is governed by a universal interaction term. Working in the Regge limit (late time $t>l_B$, small regulator) and taking the replica limit $n\\to 1$ of the six-point function, the authors obtain\n$$\\$\\Delta$ S_A[O_A;O_B]=\\$\\Delta$ S_A[O_A]+\\$\\Delta$ S_A[O_B]+\\frac{c}{6}\\log\\frac{l_B-l_A}{t-l_A}.$$\nThe interaction term is negative, exact in the central charge, and independent of how heavy the operators are. Equivalently, at very late times the double-excitation entropy grows as $\\frac{c}{6}\\log t$, not as $2\\times\\frac{c}{6}\\log t$. The authors interpret the negative term as the CFT dual of the attractive gravitational force between two particles, and they contrast this with RCFTs and free scalars, where the sum rule survives.","pith_inferences":["If Eq. (25) is correct, the same six-point block limit should control other entanglement probes in double-quench states, such as reflected entropy or negativity; computing those would test whether the interaction term is universal beyond the single-interval entropy.","Because the interaction term is independent of operator weights, one can test Eq. (25) numerically in a specific pure CFT without choosing particular operators—a mismatch would localize exactly which step in the derivation fails.","The gravitational interpretation could be sharpened by computing holographic entanglement entropy for two massive particles that fall without merging; the paper notes the merger geometry gives the right $\\log t$ growth but not the constant terms, so a non-merging two-particle geometry is a concrete candidate for a quantitative match."],"forward_implications":["In any pure CFT, the entanglement entropy after two local excitations is strictly smaller than the sum of the two single-excitation entropies, and the deficit grows logarithmically with the separation $l_B-l_A$ and with time.","At late times the double-excitation state behaves like a single effective excitation, with $\\Delta S_A\\sim \\frac{c}{6}\\log t$ rather than $2\\times\\frac{c}{6}\\log t$, consistent with two particles merging gravitationally.","The interaction term is universal: it depends only on the central charge and the distances $l_A,l_B$, not on which operators are excited.","In rational CFTs and free massless scalars the sum rule survives, so the appearance of the negative interaction term is a sharp diagnostic for theories with a holographic (Einstein-gravity) dual.","The same method yields explicit generalizations for a finite interval and for a circle (Eqs. (26) and (27)), giving concrete predictions for how the interaction term depends on interval length and total circumference."],"supporting_citations":[{"why":"Companion paper that develops the fusion/monodromy matrix method and states the unproved $n\\to1$ six-point block factorization (Eq. (23)) on which the main result rests.","marker":"[1]"},{"why":"Supplies the twist-operator correlator formula for entanglement entropy after local operator excitation, which Eq. (21) extends to two operators.","marker":"[7]"},{"why":"Demonstrates the entanglement sum rule in RCFTs, the contrast case that makes the negative interaction term specific to pure CFTs.","marker":"[10]"},{"why":"Provides the falling-particle gravity dual of a single local quench and the associated single-excitation entanglement entropy used as the sum-rule baseline.","marker":"[11]"},{"why":"Gives the holographic local-quench entanglement entropy and the heavy-state interpretation used to motivate the two-particle gravity picture.","marker":"[12]"},{"why":"Previously established the sum rule for entanglement entropy with multiple local excitations in rational and free theories, the behavior that the new interaction term corrects.","marker":"[18]"},{"why":"Earlier double local quench study attributing a similar negative correction to gravitational attraction; the paper's interpretation follows this.","marker":"[27]"}],"fun_headline_variants":["Double quench in CFTs breaks entropy sum rule","Entanglement after twin quench: less than sum of parts","Gravity's fingerprint: negative entanglement term in CFT","Sum rule fails: double quench entropy in 2D CFTs","Six-point function yields negative entanglement correction"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The result stands on two unproved inputs: that the CFT is pure (no extra conserved currents), and that in the $n\\to1$ replica limit the leftover six-point conformal block is exactly $[z_1(z_2-z_1)]^{-2h_{\\sigma_n}}$ as stated in Eq. (23), a factorization inherited from the companion paper; if either fails, the interaction term in Eq. (25) would change.","fun_headline_variants_meta":{"raw":{"variants":["Double quench in CFTs breaks entropy sum rule","Entanglement after twin quench: less than sum of parts","Gravity's fingerprint: negative entanglement term in CFT","Sum rule fails: double quench entropy in 2D CFTs","Six-point function yields negative entanglement correction"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000201,"raw_usage":{"total_tokens":1348,"prompt_tokens":883,"completion_tokens":465,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":499,"completion_tokens_details":{"reasoning_tokens":397}},"tokens_in":499,"tokens_out":465,"duration_ms":4757,"temperature":1.0,"reasoning_tokens":397,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T14:16:07.772630+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Numerically evaluate the six-point function in Eq. (21) for a specific pure CFT at finite replica number $n$ and take the $n\\to1$ limit using available conformal-block numerics; if the leftover factor is not $[z_1(z_2-z_1)]^{-2h_{\\sigma_n}}$, the predicted interaction term $\\frac{c}{6}\\log\\frac{l_B-l_A}{t-l_A}$ would be modified. A simpler check is to measure the late-time slope of $\\Delta S_A[O_A;O_B]$: Eq. (25) predicts a universal $\\frac{c}{6}\\log t$ growth, whereas the naive sum rule predicts twice that slope.","supporting_citations":[{"cited_title":"Notes on Quantum Entanglement of Local Operators","cited_arxiv_id":"1405.5875","evidence_quote":"Supplies the twist-operator correlator formula for entanglement entropy after local operator excitation, which Eq. (21) extends to two operators."},{"cited_title":"Scattering effect on entanglement propagation in RCFTs","cited_arxiv_id":"1610.06181","evidence_quote":"Demonstrates the entanglement sum rule in RCFTs, the contrast case that makes the negative interaction term specific to pure CFTs."}],"review_version":1}