{"id":"0af67033-d50d-4b05-9b55-3c5fd2781d6e","arxiv_id":"1908.01123","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The 0-magnon and 1-magnon Regge trajectories are exact functions of the fishnet coupling; 2-magnon trajectories are obtained perturbatively, with Mellin amplitudes evaluated at weak and strong coupling.","lead":"Researchers studied very high-energy limits of correlation functions in an exactly solvable toy model called conformal fishnet theory, computing so-called Regge trajectories and scattering-like amplitudes. The paper is useful as a check of Conformal Regge Theory and as a source of new non-perturbative data in a solvable CFT.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The exact Regge trajectories are algebraically sound, but the Mellin amplitudes rely on the unproved real-t assumption in Appendices A.1/A.2; non-unitary fishnet allows complex t-channel dimensions, which would invalidate the finite-ν reductions (4.8) and (5.8).","rationale":"The reader's weakest assumption identifies the same load-bearing step: the reduction of the ν integral to finite intervals in the Mellin amplitudes. The exact Regge trajectories themselves are obtained by solving algebraic equations, e.g. (4.3) and (5.3), so I do not see an internal flaw in that part of the paper. The vulnerable point is the contour manipulation in Appendices A.1 and A.2, which requires the physical spectrum of t to be real in order for products Γ(p+iq)Γ(p−iq) to be real and for the tail integrals to vanish after Wick rotation. This is explicitly stated as crucial but not proved. Because fishnet is non-unitary, complex t-channel dimensions are a plausible possibility, and the paper's own results include negative and imaginary intercepts. The same appendices also neglect lower-half-plane residue contributions with an argument that is only sketched for weak coupling. The paper itself flags a related unresolved issue in §4.2.1, where terms proportional to polygamma functions of t are discarded as a conjecture. Therefore the exact trajectories may stand, but the claimed Mellin amplitudes are conditional on an unverified analyticity assumption. This supports the reader's CONDITIONAL verdict without changing it.","tokens_in":34987,"tokens_out":21994,"duration_ms":221190,"concrete_test":"From the exact correlators of [2], extract the t-channel OPE pole positions of G0 and G1, i.e. the exchanged operator dimensions appearing in the Mellin amplitudes. If any pole has Im(Δ)≠0, the real-t premise of Appendices A.1/A.2 is false and the reductions (4.8)/(5.8) are invalid. If all exchanged dimensions are real, the premise is at least consistent with the physical spectrum; then proceed to numerically verify identity (A.6) at a sample real t and small f to confirm that the finite-interval reduction reproduces the full tail integral up to the claimed O(s^{−1}) corrections.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Equations (4.8) and (5.8) reduce the ν integral to |ν|<f² or |ν|<g using the appendix identities (A.6)/(A.37). Both derivations explicitly invoke the statement that the physical spectrum of t consists of real values only, which makes products Γ(p+iq)Γ(p−iq) real and causes the Wick-rotated tail integrals over [f²,∞) or [g,∞) to vanish. This premise is not proved. Fishnet is non-unitary, and the paper itself finds negative and even imaginary Regge intercepts, so complex t-channel operator dimensions are a priori allowed; if any such dimension exists, the Γ-product reality step fails and the tail integrals do not vanish. Separately, the same appendices discard lower-half-plane pole residues as O(s^{−1}) with only a weak-coupling plausibility argument; if any residue grows with s, the reduced amplitudes miss a non-negligible contribution. The exact trajectories (4.4) and (1.14) survive because they come from algebraic pole solving, but the claimed Mellin amplitudes (4.13), (5.15), (5.35), (5.20), and (5.40) are conditional on this reduction. The paper's own conjecture in §4.2.1 that terms proportional to polygamma functions of t must be discarded reinforces that the 0-magnon amplitude, at least, is not fully under control.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the Regge limit of Mellin amplitudes for 0-, 1-, and 2-magnon four-point correlators of the bi-scalar fishnet CFT. Using the conformal Regge theory framework of Costa-Goncalves-Penedones and the exact spectral functions of fishnet correlators, the authors solve the pole equations for the Regge trajectories. For the 0-magnon correlator the trajectories are J±2,4 = -1 ± sqrt(1-ν²) ± 2 sqrt(f⁴-ν²), and for the 1-magnon even- and odd-spin cases they are -1 ± sqrt(g²-ν²) and -1 ± i sqrt(g²+ν²), respectively. The authors then evaluate the ν integral by reducing it to finite intervals, obtaining weak- and strong-coupling Mellin amplitudes. For the 2-magnon correlator they present separate weak- and strong-coupling perturbative expansions. The 0-magnon result is compared with Korchemsky's momentum-space computation, with agreement claimed only after removing certain t-dependent polygamma terms by a conjectural prescription.","tokens_in":35248,"tokens_out":6392,"duration_ms":65892,"significance":"The exact, coupling-dependent Regge trajectories for the 0- and 1-magnon correlators are the strongest contribution of the paper: they follow from algebraic pole solving of spectral functions taken from prior independent work, with no free parameters, and they predict explicit s^{J(ν)} power laws. The extension to 1- and 2-magnon operators, for which LSZ reduction is not available, is a valuable exploitation of Mellin-space conformal Regge theory. The 2-magnon expansions, although perturbative, organize a non-trivial spectral function and expose an interesting scale splitting in ν. However, the Mellin-amplitude results are conditional on unproved contour manipulations in Appendix A and on a conjectural comparison prescription in Section 4.2.1, so the significance of the amplitude computations is not yet fully established.","major_comments":[{"comment":"The reduction of the ν-integral to the finite intervals |ν|≤f² and |ν|≤g relies on the assertion that 'the physical spectrum of t consists of real values only' (Appendix A.1 before Eq. (A.16), and Appendix A.2 before Eq. (A.29)). This assertion is not proved. Fishnet theory is non-unitary, and the paper itself finds negative and even imaginary Regge intercepts, so complex t-channel operator dimensions are a priori allowed. If any t-channel dimension has non-zero imaginary part, the products Γ(p+iq)Γ(p−iq) are not real, the tail integrals over [f²,∞) and [g,∞) do not obviously vanish after the Wick rotation, and the simplified amplitudes (4.13), (5.15), (5.20), (5.35), and (5.40) are not justified. Because these Mellin amplitudes are a central claimed result, this is a load-bearing gap. The authors should either prove the reality of the relevant physical t spectrum or clearly reformulate the amplitude results as conditional on that assumption.","section":"Appendix A.1/A.2; Eqs. (4.8), (5.8)"},{"comment":"The lower-half-plane pole residues are discarded as O(s^{-1}) contributions, with the statement that this holds 'unanimously in the weak coupling regime' (Appendix A.1). No proof or quantitative bound is given. For the 0-magnon case the trajectories J±2 have complex values for |ν|>f² and grow with f², so the accompanying Γ-functions and sin(πJ) factors can in principle produce residues whose s-dependence is not uniformly subleading. This point also affects the 1-magnon reductions in Appendix A.2. Since the same residue discarding is used to derive (4.8) and (5.8), the weak-coupling Mellin amplitudes inherit this gap. The manuscript needs either a controlled estimate of the discarded residue sums or a demonstration that they are subleading in the precise limits taken.","section":"Appendix A.1/A.2; Eqs. (A.12), (A.23)"},{"comment":"The claimed match of the 0-magnon Mellin amplitude with the momentum-space result of [3] requires discarding all terms proportional to (ψ^{(n)}(1−t/2))^m and also changing the ν-normalization convention. The discarding is presented as a conjecture with 'no deeper understanding' (Section 4.2.1). This means the comparison with [3] is not an independent confirmation of (4.13); rather, it is a conditional statement that (4.13) reduces to the known result after a non-derivative subtraction. Because the t-dependent polygamma terms are essential parts of the Mellin-space integrand, the status of (4.13) as the Regge Mellin amplitude is not fully under control. The authors should either derive the subtraction prescription or explicitly frame (4.13) as a Mellin-space result that differs from the momentum-space amplitude by those terms.","section":"Section 4.2.1; Eqs. (4.12), (4.13)"},{"comment":"The 2-magnon results rely on two unproven perturbative assumptions: the ansatz J = Σ a_n ξ^{4n/3} in the regime |ν|≤ξ⁴ (Section 6.1.2), justified only by an observation at ν=0, and the strong-coupling statement that the ν-integral is dominated by ν=0 while 'neglecting the effect of the poles' (Section 6.2.3). No error estimates or checks from independent regimes are provided. Since the displayed Mellin amplitudes (1.21) and (1.23) are central outputs of Section 6, these assumptions are load-bearing, even though the section is explicitly perturbative. The authors should state the expected size of the neglected terms or verify the ansatz by a consistency condition such as matching the two perturbative branches at |ν|∼ξ⁴.","section":"Section 6.1.2 and Section 6.2.3; Eqs. (1.21), (1.23)"}],"minor_comments":[{"comment":"There is a typo: 'Bethe-Salpater' should be 'Bethe-Salpeter'.","section":"Section 2"},{"comment":"The expression contains the typographical artifact 'Γ(J1iν+1 2)' which should presumably be Γ((J+iν+1)/2); please correct it.","section":"Eq. (5.1)"},{"comment":"The notation LLL for the modified Struve function is used in the main results before it is defined in Appendix B. Please define it at first use.","section":"Eq. (1.12) and Section 4.2"},{"comment":"The horizontal axis is labeled 'iν', which is confusing because ν is the real integration variable in the text. Please relabel the axis (e.g., as ν) or explain the intended variable.","section":"Figure 2"},{"comment":"Reference [9] is a duplicate of reference [2]; the same title is listed twice, which makes it harder to attribute the 0-magnon trajectory result.","section":"References"},{"comment":"The text refers to 'we can write (4.47) as', but no Eq. (4.47) exists in Section 4; the intended cross-reference appears to be to Eq. (5.27).","section":"Section 5.2.1"}],"recommendation":"major_revision","confidential_remarks":"The exact-trajectory part of the paper is a solid, potentially publishable contribution, and I would encourage the authors to restructure the presentation so that the unproved contour reductions are either proved or cleanly separated from the trajectory results. Because the advertised match with [3] involves a conjectural subtraction, the paper's claims about the 0-magnon Mellin amplitude should be softened accordingly. My major-revision recommendation is driven by these load-bearing gaps rather than by any doubt about the algebraic pole derivations."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The 1-magnon part is the reason to care about this paper. The even- and odd-spin trajectories J_e = -1 ± sqrt(g^2 - nu^2) and J_o = -1 ± i sqrt(g^2 + nu^2) are exact, new, and follow directly from solving the pole equations of known spectral functions. As far as I can tell, the algebra is sound, and the weak-coupling Mellin amplitudes are a legitimate new result. The 0-magnon section is a useful Mellin-space re-derivation of Korchemsky's momentum-space result, and the 2-magnon section is honestly labeled as perturbative and, in strong coupling, order-of-magnitude. I believe the central logic holds up.\n\nThe soft spots are real but localized. The reductions of the nu-integrals to the finite intervals |nu| < f^2 and |nu| < g in Appendices A.1 and A.2 lean on a statement that the physical spectrum of t consists of real values only. That makes Gamma(p+iq)Gamma(p-iq) real and kills the tail integrals, but the paper never proves it. Fishnet is non-unitary, and the paper itself finds negative and even imaginary Regge intercepts, so complex t-channel dimensions are not obviously excluded. If any such dimension exists, (4.8) and (5.8) fail. The exact trajectories survive, but the claimed Mellin amplitudes are conditional on this assumption. The same appendices discard lower-half-plane pole residues as O(s^{-1}) with only a weak-coupling plausibility argument. A referee should push on this.\n\nA second, separate caveat: the 0-magnon comparison with [3] requires dropping the polygamma-of-t terms, a step the authors themselves label a conjecture they don't fully understand. That is honest, but it means the t-dependence of the 0-magnon amplitude is not fully under control. Minor issue: references [2] and [9] are the same paper with the same arXiv number; that should be cleaned up.\n\nWho is this for? People working on integrable CFTs, fishnet theory, and Conformal Regge Theory. The exact trajectories give a useful non-perturbative data point. The paper deserves a serious referee: the main results are likely correct, but the real-t assumption needs proof or at least prominent advertisement in the abstract. I would send it to review with a request for substantial revision.","headline":"The 1-magnon Regge trajectories are the real news and are probably right, but the Mellin amplitudes rest on an unproved real-t assumption that a revision should confront.","tokens_in":35879,"tokens_out":2120,"would_cite":true,"duration_ms":23358,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["11.25.Hf","11.55.Jy"],"model":"deepseek-v4-flash","headline":"This paper derives exact Regge trajectories for the 0- and 1-magnon fishnet correlators and perturbative trajectories for the 2-magnon correlator.","keywords":["fishnet CFT","Regge limit","Mellin amplitudes","conformal Regge theory","magnon correlators","integrability","Sommerfeld-Watson transform","Regge trajectories"],"falsifier":"Evaluate the full $\\nu$ integral (4.6) numerically for real $t$ and finite $f$ without truncating to $|\\nu|<f^2$, and compare with (4.8); if tail contributions fail to vanish or the Gamma products develop complex phases after Wick rotation, the finite-interval reduction is false. Equivalently, check whether the Regge poles $J(\\nu)$ cross the Sommerfeld-Watson contour before the strong-coupling regime.","tokens_in":34690,"feed_emoji":"📐","tokens_out":8507,"duration_ms":78323,"temperature":0.7,"pith_summary":"This paper claims that in the four-dimensional conformal fishnet theory, the Regge trajectories of the 0- and 1-magnon correlators can be written down exactly as functions of the coupling, and that the corresponding Mellin amplitudes can then be evaluated in the weak- and strong-coupling limits. For the 2-magnon correlator, the claim is that only perturbative control is available, with separate expansions in different regions of the spectral parameter $\\nu$. The interest is that exact Regge data in an interacting conformal field theory is rare, and the fishnet theory's integrability makes these trajectories fully determined rather than asymptotic. If correct, these results give a concrete window into the high-energy behavior of a non-unitary CFT.","feed_headline":"Fishnet CFT yields exact Regge trajectories for magnons","feed_subtitle":"Closed-form Regge poles for 0- and 1-magnon correlators, with weak- and strong-coupling Mellin amplitudes to match.","key_machinery":"The carrying object is the Mellin amplitude in the principal-series representation, combined with the Sommerfeld-Watson transform that trades the spin sum for a contour integral in $J$. The spectral weight $b_J(\\nu^2)$ is known exactly from fishnet integrability, and the Regge trajectories are the zeros of the denominator $1 - \\chi_n E^{(n)}_{\\Delta,J}$ after setting $\\Delta = 2 + i\\nu$; for 0-magnon this denominator is $(J^2+\\nu^2)((J+2)^2+\\nu^2) - 4f^4$. The poles of the spectral function select the trajectories, and the remaining $\\nu$ integral is reduced to the finite interval $|\\nu| < f^2$ (or $|\\nu| < g$) by a contour rotation that the paper argues discards only $O(s^{-1})$ tails. This machinery converts exact correlators into explicit Regge-limit Mellin amplitudes.","core_discovery":"The central claim is that the Regge pole locations of the Mellin amplitudes are fixed by the poles of the exactly known spectral function, so the trajectories are exact functions of the coupling. For the 0-magnon correlator the paper obtains $J^\\pm_2(\\nu) = -1 + \\sqrt{1-\\nu^2} \\pm 2\\sqrt{f^4-\\nu^2}$ (leading trajectory $J^+_2$) and $J^\\pm_4(\\nu) = -1 - \\sqrt{1-\\nu^2} \\pm 2\\sqrt{f^4-\\nu^2}$. For the 1-magnon correlator it obtains $J^\\pm_e = -1 \\pm \\sqrt{g^2-\\nu^2}$ for even spin and $J^\\pm_o = -1 \\pm i\\sqrt{g^2+\\nu^2}$ for odd spin. For the 2-magnon correlator no all-coupling closed form is found; instead the paper derives weak-coupling expansions in two $\\nu$ regimes and a strong-coupling expansion in powers of $1/\\xi$. The resulting Regge-limit Mellin amplitudes are evaluated at weak coupling in terms of Bessel and Struve functions, and at strong coupling as power laws with trigonometric prefactors.","pith_inferences":["Extension: if the exact trajectories hold, they provide a benchmark for numerical bootstrap studies of fishnet-type CFTs, since Regge intercepts are sensitive to the full spectrum.","Extension: the same machinery applies to any integrable CFT with an exactly known spectral function; the input needed is only the denominator $1-\\chi E(\\nu,J)$ of the graph-building operator.","Extension: the t-dependent discrepancy between Mellin-space and momentum-space results suggests a subtraction dictionary involving $\\psi^{(n)}(1-t/2)$ terms, which could be tested on the 1- and 2-magnon correlators.","Extension: evaluating the exact $\\nu$ integral numerically at intermediate coupling would test whether $J^+_2(\\nu)$ remains the leading Regge pole for all couplings or whether another trajectory crosses it."],"forward_implications":["The 0-, 1-, and 2-magnon leading intercepts at weak coupling are 0, -1, and -2, matching the non-unitary exchanges expected from fishnet theory.","Strong-coupling amplitudes carry factors $\\csc(\\sqrt{2}\\pi f)$ and $\\csc(\\pi g)$, so the Regge amplitudes develop periodic singularities in the coupling; these are explicit predictions for where the saddle-point approximation breaks down.","The 2-magnon analysis shows two distinct weak-coupling regimes separated at $|\\nu| \\sim \\xi^4$, with different analytic forms for the trajectory; this is a level-crossing structure that a full all-coupling solution would have to reproduce.","The finite-interval reduction means the entire leading Regge amplitude in the weak-coupling limit is controlled by a one-dimensional integral with explicitly known Bessel and Struve kernels, allowing systematic higher-order corrections."],"supporting_citations":[{"why":"Supplies the Mellin-space conformal Regge formalism and the principal-series representation of the amplitude.","marker":"[7]"},{"why":"Provides the momentum-space Regge amplitude for 0-magnon and the contour-reduction trick adapted here.","marker":"[3]"},{"why":"Gives the exact n-magnon correlators and the graph-building eigenvalues $E^{(n)}_{\\Delta,J}$ used as input.","marker":"[2]"},{"why":"Fixes the exact spectral function $b_J(\\nu^2)$ for the fishnet CFT, the key exact input.","marker":"[8]"},{"why":"Defines the bi-scalar fishnet CFT and its double-scaling limit.","marker":"[1]"}],"fun_headline_variants":["Exact Regge trajectories for 0- and 1-magnon correlators","Regge limit of fishnet correlators: exact poles for 0 and 1 magnons","Fishnet Regge trajectories: exact for low magnons","Exact Regge poles for low-magnon fishnet correlators","0- and 1-magnon Regge poles exact; 2-magnon perturbative"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The reduction to the finite $\\nu$ interval assumes the physical spectrum of $t$ consists of real values only, so products $\\Gamma(p+iq)\\Gamma(p-iq)$ are real and the Wick-rotated tail integrals vanish; if $t$ can take complex values, or if the Regge poles migrate to the wrong half-plane at finite coupling, the simplified integrals (4.8) and (5.8) are not justified.","fun_headline_variants_meta":{"raw":{"variants":["Exact Regge trajectories for 0- and 1-magnon correlators","Regge limit of fishnet correlators: exact poles for 0 and 1 magnons","Fishnet Regge trajectories: exact for low magnons","Exact Regge poles for low-magnon fishnet correlators","0- and 1-magnon Regge poles exact; 2-magnon perturbative"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001086,"raw_usage":{"total_tokens":4521,"prompt_tokens":908,"completion_tokens":3613,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":524,"completion_tokens_details":{"reasoning_tokens":3506}},"tokens_in":524,"tokens_out":3613,"duration_ms":25952,"temperature":1.0,"reasoning_tokens":3506,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T15:23:15.395168+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Evaluate the full $\\nu$ integral (4.6) numerically for real $t$ and finite $f$ without truncating to $|\\nu|<f^2$, and compare with (4.8); if tail contributions fail to vanish or the Gamma products develop complex phases after Wick rotation, the finite-interval reduction is false. Equivalently, check whether the Regge poles $J(\\nu)$ cross the Sommerfeld-Watson contour before the strong-coupling regime.","supporting_citations":[],"review_version":1}