{"id":"65683671-d8f9-4066-ad89-c6492ab649f8","arxiv_id":"2411.18696","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"The bulk energy four-point function in percolation and self-avoiding walk CFTs is non-zero at c=0, driven by coupling to a rank-3 Jordan block associated with the second energy operator.","lead":"This paper studies Kac operators in two-dimensional conformal field theories at central charge c=0, the theories that describe percolation and self-avoiding walks. It finds that the four-point function of the energy operator does not vanish, contrary to earlier belief, because the operator couples to a higher-rank logarithmic structure.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Four-point function rests on unproven cluster decomposition at c=0 and is in direct tension with the vanishing c→0 limit of the BPZ four-point function.","rationale":"The reader's weakest-assumption identification (cluster decomposition at c=0) is exactly the load-bearing point of the central claim. I agree, and I sharpen it in two ways. First, the paper's own eq. (5.49) shows that the naive c→0 limit of the BPZ four-point function vanishes at leading order; the nonzero result (5.55) is therefore not a limit of the generic-c theory but an additional postulate about the exact c=0 logarithmic theory. Whether this postulate is valid is precisely the question of whether cluster decomposition and the completeness of the state set (5.51) hold in a non-unitary logarithmic CFT. The paper offers only 'all conformal dimensions are positive' as justification, which is not sufficient because cluster decomposition in such theories is not a trivial consequence of spectral positivity. Second, the numerical coefficient is tied to the chosen normalization B_ε ≃ c/2; while the existence of a nonzero four-point function is scale-invariant, the specific value in (5.55) is not, so even if the check passes the numerical agreement would require fixing this normalization independently. The proposed concrete check—a crossing-symmetric logarithmic bootstrap at c=0—directly tests the consistency of the OPE data used to assemble (5.55). If no crossing-symmetric solution exists, the cluster-decomposition computation cannot be a correlation function of a consistent CFT; if it exists, the central claim is supported. This is a decisive, analytic test that does not require interpreting subtle lattice operator identifications. I therefore leave the verdict as CONDITIONAL, matching the reader's assessment.","tokens_in":64159,"tokens_out":16918,"duration_ms":159931,"concrete_test":"Run the logarithmic conformal bootstrap at c=0 for ⟨εεεε⟩: impose crossing symmetry on a four-point function built from the spectrum and three-point couplings of §5.1 (rank-2 block (t,T), rank-3 block (Ψ2,Ψ1,Ψ0), vanishing two-point function of ε, and the constants of eq. (5.21)). Determine whether a crossing-symmetric solution exists whose s-channel limit reproduces eq. (5.55). If the crossing equations force the four-point function to vanish, the cluster-decomposition assumption is inconsistent; if a unique crossing-symmetric solution exists, the claim is corroborated. A complementary lattice check: measure the four-point correlation of the bond-occupation energy in critical percolation at the geometry of two coalescing pairs and test whether the leading term scales as (|r1-r2||r3-r4|)^{3/4} with a nonzero coefficient.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central result, eq. (5.55), is obtained by assembling the four-point function of the zero-norm energy operator ε from three-point functions via cluster decomposition, eq. (5.50), using the finite state list (5.51) and the inverse Gram matrix. This assumes that at c=0 the state space admits a complete set of intermediate states and that the usual OPE/conformal-block expansion survives in a non-unitary logarithmic CFT whose vacuum module contains zero-norm states and Jordan blocks. The paper's justification ('all conformal dimensions are positive', §5.3) is not sufficient: cluster decomposition is not guaranteed by positivity of dimensions in a logarithmic theory, and the derivation of (5.55) requires not merely positivity but the completeness of the specifically chosen set {ψ} up to order (z z̄)². The tension is concrete: the direct c→0 limit of the generic-c BPZ four-point function, eq. (5.42)–(5.49), vanishes at O(1) in the s-channel, with only O(c) logarithmic terms surviving. The nonzero result (5.55) therefore does not follow from the limit of the generic-c theory; it is a new assumption about the exact c=0 theory, namely that the middle-field Ψ1 of the rank-3 Jordan block couples to ε with the finite constant (5.11) and that cluster decomposition applies. If cluster decomposition fails, or if additional intermediate states (e.g. other primaries at dimension (2,2)) contribute, the four-point function could vanish or acquire a different form. The numerical coefficient (C²/a ≈ -0.114) is also normalization-dependent, since the choice B_ε ≃ c/2 in eq. (4.13) fixes the overall scale of the three-point constant and hence of the four-point function; a different positive prefactor k in B_ε ~ k c/2 multiplies (5.55) by k⁴. The non-vanishing statement is robust to this rescaling, but the precise prediction is not.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies logarithmic operators in c=0 bulk CFTs describing percolation and self-avoiding walks. It first revisits the known rank-2 and rank-3 Jordan blocks involving T and T-bar-T, then proposes proper normalizations for Kac operators in cluster and loop models at generic c, using analytic bootstrap amplitudes and the requirement that three-point constants be real and finite. At c=0 the Kac operators acquire zero norms, and the paper constructs logarithmic partners and computes c=0 conformal data, including the logarithmic couplings b=-5 and a=-25/48, the three-point constant C^perco_εεΨ1=-125/512, and related OPE coefficients. The central claim, stated in Section 5.3 and summarized in the abstract, is that the four-point function of the bulk energy operator does not vanish at c=0, with the leading s-channel term given in eq. (5.55) as (C^perco_εεΨ1)^2/a (z zbar)^2, built through coupling to the middle field Ψ1 of the rank-3 Jordan block. The paper closes with a discussion of the puzzling physical interpretation and suggestions for future checks.","tokens_in":64493,"tokens_out":6984,"duration_ms":73938,"significance":"If the central claim is correct, the paper establishes an intriguing mechanism by which zero-norm Kac operators can generate long-range higher-point correlations through logarithmic Jordan-block structures, overturning the earlier belief that such correlators vanish at c=0. The work contains substantial strengths: the operator normalizations are deduced from parameter-free amplitude recursions of the analytic bootstrap; the c=0 data are obtained by well-defined c→0 limits with many internal consistency checks; the known values b=-5 and a=-25/48 are recovered; and the results satisfy conformal Ward identities as summarized in Appendix D. The comparison with c<1 Liouville CFT normalizations in Section 3.4 is also suggestive and well presented. However, the headline four-point result is explicitly conditional on an unproven cluster-decomposition assumption at c=0, and it is in direct tension with the vanishing O(1) term of the direct c→0 limit of the BPZ four-point function.","major_comments":[{"comment":"The central non-vanishing four-point function is assembled by inserting a complete set of intermediate states {ψ} in (5.51) and inverting the Gram matrix. This requires cluster decomposition and completeness of that specific set at c=0, but the only justification offered is the sentence before (5.50) that all conformal dimensions are positive. Positivity of dimensions does not imply cluster decomposition in a non-unitary logarithmic theory with zero-norm states and Jordan blocks, and the finite list in (5.51) is asserted rather than derived; other states with dimensions up to (2,2) could contribute to the O(z zbar)^2 term. The load-bearing character of this assumption is underscored by the direct c→0 limit of the BPZ four-point function, eqs. (5.42)-(5.49), which vanishes at O(1) and leaves only O(c) logarithmic terms. The nonzero O(1) result (5.55) is therefore not obtained as a limit of the generic-c correlator; it depends on an additional assumption about the exact c=0 theory. Please either prove cluster decomposition and the completeness of (5.51) in this logarithmic setting, or provide an independent lattice or probabilistic computation of the four-energy correlator or of the coupling ⟨εεΨ1⟩ that confirms (5.55).","section":"5.3 (eqs. (5.50)-(5.55))"},{"comment":"The central coefficient C^perco_εεΨ1=-125/512 and hence the four-point amplitude (5.55) depend on the normalization choice B'_ε=1/2 in eq. (4.13). This choice is motivated by analogy with the stress tensor, but the energy operator is not the stress tensor and no symmetry forces this value. The comparison with [6] in Section 4.2.1 fixes some product of normalization constants, but the paper should state explicitly that the non-vanishing four-point function is contingent on the lattice-derived normalization of the energy operator. A different positive normalization would rescale the four-point function, and a negative B'_ε, which is not excluded by the reality arguments in Section 3.1, could change its sign and potentially the conclusion. Please derive the normalization from an independent physical definition of the energy operator in the cluster/loop model, or clearly delimit the dependence of (5.55) on this choice.","section":"4.2 (eq. (4.13))"},{"comment":"The derivation of the three-point coupling C^perco_εεΨ1 uses the asymptotic expansion C_{Φ2,1Φ2,1Φ3,1}(c) ≃ h_ε^2 c/2 + ... (footnote 18) and similar singularity-cancellation conditions in Appendix B.2. These expansion coefficients are load-bearing inputs: if the O(c) term or the subleading coefficient differs, the finite three-point constant (5.11) changes and the four-point result (5.55) changes with it. The paper states only that these conditions 'can be checked,' without displaying the expansions or their derivations. Please present the explicit c-expansions and the verification of conditions (B.20)-(B.21) in the main text or appendix.","section":"5.2 (eqs. (5.10)-(5.11))"},{"comment":"The reconciliation with the BPZ limit is interpretive rather than derivational. The paper argues that the generic-c Kac operator 'only knows about' the intermediate rank-2 block (Θ, Φ̂3,1) and not about Ψ1, but if the exact c=0 theory is a limit of the generic-c family, the four-point function should equal the limit (5.49) unless the c→0 limit and the cluster-decomposition sum fail to commute. The paper does not prove such non-commutation, nor does it define the exact c=0 theory by an independent set of axioms that would make the enlarged intermediate-state space (5.51) self-consistent. Please make this logical step explicit and provide a concrete definition of the c=0 theory under which (5.50) is the correct expansion.","section":"5.3 (eqs. (5.56)-(5.57))"}],"minor_comments":[{"comment":"There is a typo in the Conclusions: 'energy operataor' should be 'energy operator'; the manuscript would benefit from a careful proofread.","section":"Section 6"},{"comment":"The displayed expression for C^{SAW}_{εεε} contains the ambiguous factor 'Γ(1/6)^{7/2}'; please clarify the placement of the exponent and the parentheses.","section":"Eq. (5.27)"},{"comment":"The notation B^{(2)}_{Φ3,1}, B^{(3)}_{Φ3,1} is used without a uniform definition; please define these as the coefficients of (c-c*)^2 and (c-c*)^3 in the expansion of the two-point constant B_{Φ3,1}(c).","section":"Eqs. (2.38), (4.47), (5.49)"},{"comment":"The sentence says 'there are five fields with dimensions (2,2)' and lists ∂^2 t̄, ∂̄^2 t, Ψ2, Ψ1, Ψ0, but the preceding discussion refers to four fields in the rank-3 block; the counting deserves a clearer statement.","section":"After eq. (2.48)"},{"comment":"The comparison with c<1 Liouville CFT is used to suggest arbitrarily high-rank Jordan blocks, but the preceding analysis in Section 3.2.1 leaves the order of zeros for Φ4,1 and Φ5,1 unresolved; please state explicitly which higher-rank conclusions are conjectural and which are established.","section":"Section 3.4"}],"recommendation":"major_revision","confidential_remarks":"The paper is ambitious and contains a large amount of correct-looking internal consistency checking, but the headline claim is conditional on an assumption that the author acknowledges. I would ask the editor to require a substantive new test before acceptance: either a proof or a precise axiomatic definition of cluster decomposition at c=0, a lattice or probabilistic computation of the four-energy correlator, or an independent derivation of the normalization in eq. (4.13). Without such a test, the paper would be more appropriately framed as a conjecture than as an established result. The scope is suitable for a hep-th/CFT journal, and the literature appears adequately cited."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper is worth taking seriously. Its real contribution is the normalization of Kac operators at generic c from the reality principle, and a clean, systematic construction of rank-2 and rank-3 Jordan blocks from vanishing norms. The recovery of known values like b = -5 and a = -25/48, the singularity cancellation checks in the appendices, and the agreement with earlier lattice and bootstrap data are solid evidence that the machinery is not just formal. The comparison with c<1 Liouville CFT is a nice touch, even if the physical explanation is still open.\n\nThe central claim—that the bulk energy four-point function does not vanish at c=0—is intriguing, and the mechanism (coupling to the rank-3 Jordan block via the middle field Ψ1) is concrete and computable. But the stress-test concern is legitimate: eq. (5.55) is assembled by inserting a finite set of intermediate states and assuming cluster decomposition at c=0. The paper states this assumption rather than proving it, and the direct c→0 limit of the BPZ four-point function vanishes at O(1), so the non-vanishing result is genuinely an assumption about the exact c=0 theory, not a consequence of the limit. The paper does acknowledge this tension in Section 5.3 and in the conclusions, but the abstract states the result more firmly than the body warrants. The numerical coefficient in (5.55) is also normalization-dependent through the choice B'_ε = 1/2; the qualitative non-vanishing statement survives rescaling, but the precise prediction does not.\n\nA second soft spot, minor in comparison, is the ambiguity in the order of zeros for higher Kac operators, which the paper honestly discusses. That caveat does not undermine the main construction for the energy operator and its second-energy partner.\n\nWho should read this: anyone working on c=0 CFTs, percolation, or SAW, and anyone interested in logarithmic operator algebras. It deserves a serious referee. A good referee should push for an independent check of the four-point function—lattice numerics or a probabilistic argument—and for a clearer statement of what exactly is assumed about completeness of intermediate states at c=0. My own verdict is conditional: the framework is credible and the machinery is well-built, but the headline result needs independent confirmation.","headline":"A technically rich and genuinely new construction of logarithmic multiplets at c=0, whose central non-vanishing energy four-point function is real but rests on an unproven cluster-decomposition assumption that the paper itself flags.","tokens_in":65099,"tokens_out":1275,"would_cite":true,"duration_ms":82253,"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":"This paper argues that zero-norm 'energy' operators in percolation and self-avoiding walk CFTs can have non-vanishing long-range four-point correlations at c=0, via a rank-3 logarithmic Jordan block.","keywords":["c=0 conformal field theory","logarithmic operators","Jordan blocks","Kac operators","percolation","self-avoiding walk","cluster decomposition","conformal bootstrap"],"falsifier":"Compute the connected four-point correlation of the local bond-occupation energy operator on critical percolation clusters in the scaling limit; in the s-channel it should behave as (z \\bar z)^{3/4} times -1875/16384 with no logarithm at leading order, rather than vanish.","tokens_in":63931,"feed_emoji":"🕸️","tokens_out":11668,"duration_ms":96594,"temperature":0.7,"pith_summary":"Percolation and self-avoiding walks are described by c=0 conformal field theories, whose 'energy' Kac operators have zero norm and were previously thought to have no long-range multipoint correlations. This paper establishes the contrary: after fixing the operators' normalizations by requiring real finite three-point constants at generic c and passing to c=0, the energy operators become bottom fields of logarithmic Jordan blocks, and the four-point function assembled by cluster decomposition is nonzero. The nonvanishing term is carried by a rank-3 Jordan block built around the second energy operator. If the argument holds, zero-norm states in c=0 bulk CFTs build long-range correlations through logarithmic structures, and the operator algebra of these geometrical critical points is richer than previously believed.","feed_headline":"Rank-3 Jordan block carries the c=0 energy four-point function","feed_subtitle":"In percolation and self-avoiding walks, the energy operator still has a nonzero four-point correlation at c=0.","key_machinery":"The machinery is the logarithmic (Jordan) multiplet formed when a Kac operator's norm vanishes at $c=0$. The paper defines properly normalized Kac operators $\\hat{\\Phi}_{r,s}$ at generic $c$, requires real and finite three-point constants, and takes the $c\\to 0$ limit. A first-order zero in the norm produces a rank-2 Jordan block $(\\tilde O,O)$; a second-order zero produces a rank-3 block $(\\Psi_2,\\Psi_1,\\Psi_0)$. The top field is built as a combination $\\phi/B_\\phi(c)+\\psi/B_\\psi(c)$ (and for rank 3, plus $O^{(2;2)}/\\gamma_O(c)$), with logarithmic couplings fixed by conformal Ward identities, for example $\\gamma = B'_\\phi/(h'_\\phi-h'_\\psi)$ and $a = \\gamma'_O/(2h'_O-h'_\\phi-h'_\\psi)$. The four-point function is then assembled by cluster decomposition over intermediate states, and the load-bearing identity is the nonzero three-point coupling $C^{\\mathrm{perco}}_{\\varepsilon\\varepsilon\\Psi_1}$ to the middle field of the rank-3 block.","core_discovery":"The paper's central discovery is that the four-point function of the bulk energy operator ε at c=0 does not vanish. For percolation, the s-channel limit is given by eq. (5.55): $$\\langle \\varepsilon(\\infty)\\varepsilon(1)\\varepsilon(z,\\bar z)\\varepsilon(0)\\rangle = (z\\bar z)^{-$2h^{{\\mathrm{perco}}$}_\\varepsilon}\\left(\\frac{($C^{{\\mathrm{perco}}$}_{\\varepsilon\\varepsilon\\Psi_1})^2}{a}(z\\bar z)^2+\\cdots\\right)$$ with $C^{\\mathrm{perco}}_{\\varepsilon\\varepsilon\\Psi_1}=-125/512$ and $a=-25/48$. The same construction applies to self-avoiding walks. The nonzero term comes from the middle field $\\Psi_1$ of the rank-3 Jordan block $(\\Psi_2,\\Psi_1,\\Psi_0)$ associated with the second energy operator $\\varepsilon'\\sim\\Phi_{3,1}$, rather than from the naive $c\\to 0$ limit of the generic-$c$ BPZ correlation, which vanishes. Cluster decomposition using the exact $c=0$ logarithmic conformal data is what reveals the surviving correlation.","pith_inferences":["If the claim is right, a direct lattice measurement of the connected four-energy correlation in critical percolation should see a nonzero long-range term, with the sharp target coefficient -1875/16384 in the leading s-channel term, something not previously looked for.","The mechanism suggests a general principle for c=0 bulk CFTs: any zero-norm Kac operator that logarithmically mixes with a hull-type operator can acquire long-range higher-point correlations through higher-rank Jordan blocks, even when its two- and three-point bottom couplings vanish.","The Liouville comparison implies that the order of zeros in Kac-operator norms may be a universal function of c; if so, the proposed rank-4 and rank-5 blocks could be tested independently by loop-model transfer-matrix or bootstrap methods, without waiting for lattice four-energy data.","A probabilistic construction of percolation correlations could give a geometric explanation of why the energy four-point function is non-logarithmic at leading order and why it is built through the rank-3 block rather than directly through bond correlations."],"forward_implications":["The energy-operator OPE at c=0 is nonsingular after proper normalization: the bottom-field couplings with T and Ψ0 vanish, while couplings to the top fields are nonzero, with C^{perco}_{εεt}=-25/8 and C^{perco}_{εεΨ1}=-125/512.","The s-channel four-energy correlation in percolation and self-avoiding walks is nonzero and, at leading order, has no logarithm, so its scaling is a power law, (z z̄)^{3/4} in percolation.","Higher Kac operators acquire higher-order zero norms at c=0, and comparison with c<1 Liouville CFT suggests rank-4, rank-5, and possibly arbitrarily high Jordan blocks, though these are invisible in the spin OPE because the corresponding C_{σσΦ_{r,1}} vanish.","The hull spin OPE only probes bottom fields, so its four-point function contains no logarithm, meaning the same Jordan blocks act differently in different sectors of the c=0 theories."],"supporting_citations":[{"why":"states the earlier conclusion that higher-point functions of bottom fields in Jordan blocks vanish, which the paper overturns.","marker":"[5]"},{"why":"argued that Kac operators at c=0 have vanishing norms and sit at the bottom of Jordan blocks, the starting premise for resolution iii of the c→0 catastrophe.","marker":"[25]"},{"why":"gives the c→0 limiting analysis whose conclusion of trivial higher-point correlations is contradicted by this paper.","marker":"[28]"},{"why":"previous discovery of the rank-3 Jordan block for T\\bar T at c=0; supplies the construction and consistency checks used here.","marker":"[21]"},{"why":"provides the analytic amplitudes of Kac operators at generic c that are used to fix operator normalizations.","marker":"[15]"},{"why":"earlier S_Q-based construction of the percolation energy pair; used to check normalizations and the logarithmic coupling.","marker":"[6]"},{"why":"supplies the principle that real CFTs have real and finite three-point constants, which is used to choose normalizations.","marker":"[26]"},{"why":"gives lattice algebraic evidence for arbitrarily high rank Jordan blocks at c=0, discussed as a consequence of higher-order zero norms.","marker":"[47]"}],"fun_headline_variants":["Energy four-point survives c=0 via rank-3 Jordan block","c=0 energy correlator nonzero from Jordan block","Zero-norm states drive long-range c=0 correlations","Rank-3 block rescues c=0 energy four-point function","Energy operator four-point not vanishing at c=0"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument assumes the four-point function can be assembled by cluster decomposition at c=0 from three-point functions, inserting only the identified intermediate states (5.51); if cluster decomposition fails or extra states contribute, the nonvanishing result could be modified.","fun_headline_variants_meta":{"raw":{"variants":["Energy four-point survives c=0 via rank-3 Jordan block","c=0 energy correlator nonzero from Jordan block","Zero-norm states drive long-range c=0 correlations","Rank-3 block rescues c=0 energy four-point function","Energy operator four-point not vanishing at c=0"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000417,"raw_usage":{"total_tokens":2195,"prompt_tokens":1036,"completion_tokens":1159,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":652,"completion_tokens_details":{"reasoning_tokens":1076}},"tokens_in":652,"tokens_out":1159,"duration_ms":9549,"temperature":1.0,"reasoning_tokens":1076,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T10:58:24.575897+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the connected four-point correlation of the local bond-occupation energy operator on critical percolation clusters in the scaling limit; in the s-channel it should behave as (z \\bar z)^{3/4} times -1875/16384 with no logarithm at leading order, rather than vanish.","supporting_citations":[{"cited_title":"The Stress Tensor in Quenched Random Systems","cited_arxiv_id":"cond-mat/0111031","evidence_quote":"argued that Kac operators at c=0 have vanishing norms and sit at the bottom of Jordan blocks, the starting premise for resolution iii of the c→0 catastrophe."},{"cited_title":"A note on the identity module in $c=0$ CFTs","cited_arxiv_id":"2109.05050","evidence_quote":"previous discovery of the rank-3 Jordan block for T\\bar T at c=0; supplies the construction and consistency checks used here."}],"review_version":1}