{"id":"bcfeb8f2-b23f-4a32-86a1-563741e1aa8a","arxiv_id":"2501.12450","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Explicit modular S and T matrices are derived for Carrollian (bms3) and warped CFT characters, with checks of S squared and unitarity.","lead":"Two physicists compute the modular S and T matrices that describe how torus characters transform under the modular group for two families of two dimensional quantum field theories with Carrollian and warped symmetries. The matrices provide a basic tool for studying black hole entropy and holography in flat and warped spacetimes.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The S-matrix derivation uses delta-function localization at complex τ, which is unjustified; without an analytic continuation the central claim holds only as a formal real-axis identity.","rationale":"The reader identifies the complex-modulus delta-function problem as the weakest assumption, and my independent reading agrees. The load-bearing step is (3.7): the ansatz fixes γ(p′,p⋆)=1/τ, but for all cases treated γ is real-valued, so this equation cannot be satisfied for Im τ>0. The authors explicitly warn that the method only works for real τ (end of Section 3), and no continuation is provided. This matters because the objects being transformed — the characters with q=e^{2πiτ} — are only defined as convergent series/functions for Im τ>0; on the real axis they are not convergent and the eta function has singularities. The formal delta manipulations in (4.17) and analogous equations treat σ as real: the Jacobian identity δ(f(P))=δ(P−P0)/|f′(P0)| presupposes a real root P0, which does not exist when σ is complex. The algebraic checks S²=1 and SS†=1 are kernel identities that do not depend on τ and thus cannot compensate. The real-charge non-unitary warped case is also left unsolved, as the paper states. These are disclosed limitations, not hidden flaws, and the final S-matrices may well be correct; but the central claim as stated — that these are the modular S-matrices generating the modular transformations on the characters — is not proven for Im τ>0. The appropriate verdict remains CONDITIONAL, so I recommend no change.","tokens_in":17219,"tokens_out":14070,"duration_ms":139342,"concrete_test":"For the Carrollian highest-weight S-matrix (4.16), set σ=i, ρ=0, and evaluate the right-hand side of (4.7) with a Gaussian regulator: R_ε = ∫ dP_L dP_M S_char(P',P) χ(P) e^{-ε(P_L²+P_M²)}, then take ε→0+. Compare with χ_{P'}(−1/σ, ρ/σ²) = e^{-2πiP'_L/σ} e^{2πiρ P'^2_M/σ²} / η(−1/σ)² (using (2.11)). If the limit does not exist, or differs by a σ-dependent factor, the delta-localization step (4.17) is not a valid analytic continuation. Alternatively, repeat the localization derivation with σ=i: since γ±(P_M,P'_M)=±P'_M/P_M is real for real P_M, equation (4.9) has no solution, demonstrating the derivation cannot be repeated for Im σ>0.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central derivations (Eqs. (4.16), (4.32), (5.12), (5.16)) all rely on the delta-function localization condition (3.7), where γ(p,p′) is real for real momenta. Since the torus modulus τ (or σ) has Im τ > 0, the equation γ(p′,p⋆)=1/τ has no solution on the real integration contour. The paper itself notes in Section 3 that if γ is real the procedure only works for real τ, but no analytic-continuation argument is supplied. The characters are convergent holomorphic functions only for Im τ > 0 (e.g., q=e^{2πiτ} with |q|<1), so an identity established only on the real axis is not a statement about the physical torus. In the explicit calculation (4.17), the step δ(σ ± P′_M/P_M)=δ(P_M ± P′_M/σ) |P′_M|/(|P_M||σ|) treats σ as a real variable; for complex σ the support lies off the real P_M axis and the standard Jacobian formula does not apply. The subsequent checks S²=1 and SS†=1 are algebraic kernel identities independent of τ and do not repair the missing modular-transformation proof. The warped cases (5.12), (5.16) suffer the same defect because γ(p,p′)=p′/p is real. Thus the strongest claim is established only formally; whether the S-matrices act as claimed on torus characters for Im τ > 0 remains unproven.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper derives modular S and T matrices for characters of two-dimensional Carrollian CFTs (bms3 algebra) in highest-weight and induced representations, and for warped CFTs in non-unitary imaginary-charge and unitary real-charge representations. The method assumes an exponential ansatz for the S kernel and uses delta-function localization to solve the defining integral equation, after which the authors verify properties such as S^2=1 for Carrollian characters and SS^dagger=1, S^2=C for warped characters. The T matrices are found to be essentially identity matrices with phases. The paper explicitly notes that its localization procedure only works for real torus modulus when the localization function is real, but proceeds to present the results without supplying an analytic-continuation argument.","tokens_in":17525,"tokens_out":5060,"duration_ms":53445,"significance":"If established rigorously, the results would provide explicit modular kernels for non-Lorentzian CFT characters, which are relevant for flat-space holography, warped AdS/CFT, and Cardy-like density-of-states computations. The paper contains several concrete algebraic checks, including explicit verification of S^2, unitarity, and charge-conjugation properties, which are valuable and go beyond a purely formal proposal. However, the central defect is that the defining modular-transformation property is only demonstrated on the real axis for the torus modulus, whereas the characters are holomorphic functions on the upper half plane. This is a load-bearing gap, not a cosmetic issue.","major_comments":[{"comment":"The derivation of the Carrollian S matrix relies on delta-function localization at a complex torus modulus, but the localization condition (3.7), gamma(p',p_star)=1/tau, has no solution on the real integration contour when gamma is real and Im tau > 0. In the explicit calculation (4.17), the step delta(sigma +/- P'_M/P_M) = delta(P_M +/- P'_M/sigma) |P'_M|/(|P_M||sigma|) treats sigma as a real variable; for Im sigma > 0 the support lies off the real P_M axis and the standard Jacobian formula does not apply. The paper itself notes in Section 3 that for real gamma the procedure only works for real tau, but no analytic-continuation argument is supplied. Since the characters are holomorphic functions on the upper half plane and are not ordinary functions on the real axis, this leaves the central claim--that (4.16) is the modular S matrix acting on torus characters--unproven for the physical domain Im sigma > 0.","section":"Section 3, Eqs. (3.6)-(3.11); Section 4.1, Eqs. (4.16)-(4.17)"},{"comment":"The same deficiency affects the warped S matrices. In both cases gamma(p_im,p'_im)=p'_im/p_im (or p'/p) is real for real momenta, so the condition (3.7) cannot be satisfied for Im tau > 0; the derivations of (5.12) and (5.16) are therefore formal real-axis computations. The subsequent checks SS^dagger=1 and S^2=C in Section 5.3 are algebraic kernel identities independent of tau and do not repair the missing modular-transformation proof on the upper half plane.","section":"Section 5.1, Eq. (5.12); Section 5.2, Eq. (5.16)"}],"minor_comments":[{"comment":"The abstract claims modular matrices for 'two dimensional Carrollian and warped CFT algebras' broadly, but Section 5.1 explains that the method fails for non-unitary warped CFTs with real charge; the summary of results should explicitly exclude that case.","section":"Abstract and Section 5.1"},{"comment":"The lower integration limit in the first double integral is printed as '∞' instead of '−∞'; the same typo appears in the surrounding line.","section":"Eq. (5.25)"},{"comment":"The notation is slightly confusing: Eq. (3.7) states gamma(p',p_star)=1/tau, and the following line multiplies by tau; it would help to state explicitly that p_star is the solution of tau=gamma(p_star,p') and therefore depends on tau and p'.","section":"Section 3, Eq. (3.7)"}],"recommendation":"major_revision","confidential_remarks":"The main gap is one that the authors themselves flag in Section 3: the delta-function localization argument is only valid for real torus modulus, and no analytic continuation is provided. This is not a case of hidden circularity, but it is load-bearing because the advertised modular transformation is a statement about characters on the upper half plane. The paper would be acceptable if the authors supply a rigorous continuation argument, or alternatively reframe the results as formal identities on a real slice with a clear statement about their physical status."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Informally: the paper gives explicit modular S and T matrices for Carrollian and warped CFT characters, which is genuinely new representation-theoretic data for those algebras. The computations are transparent and the authors check the defining properties. That part is solid.\n\nWhat is actually new: Eqs. (4.16), (4.32), (5.12), and (5.16) don't appear in the cited literature. The method – solving the integral equation with an exponential ansatz and then using delta-function localization – is a reasonable extension of the Gaussian trick used for Virasoro characters. The paper also checks S^2 = 1 for the Carrollian case and SS† = 1, S^2 = C for the warped cases. The authors are honest about the method's limitations: it fails for non-unitary warped CFTs with real charge, and they say so.\n\nThe main soft spot is the one the stress-test flags. The localization step requires solving γ(p,p') = τ for p', and when γ is real (as it is throughout), this can only have a solution on the real τ line. The authors note this in Section 3 but then proceed anyway, and no analytic continuation is supplied. Since the characters are holomorphic only for Im τ > 0, the derivation as written establishes the modular-transformation identity only formally. That's not a hidden flaw – it's disclosed – but it does mean the central claim is not fully proven for the physical regime. A direct check that the integral of the S-kernel against the character equals χ(-1/τ) for Im τ > 0 would close the gap. The algebraic checks of S^2 and unitarity are independent of τ and don't repair that.\n\nThere is also a minor gap: the derivation produces S matrices that satisfy some defining properties, but it doesn't address uniqueness. The ansatz is chosen ad hoc, so the paper establishes existence of a candidate, not that it is the unique modular matrix. That's worth a sentence in the discussion, but it's a minor point.\n\nWho this is for: people working on Carrollian CFTs, flat holography, or warped CFTs. The companion paper using these matrices for spectral densities is a natural follow-up. This is a useful subfield contribution, not a paradigm shift.\n\nMy recommendation: send it to peer review. A referee should ask for either an analytic-continuation argument or a direct integral check for Im τ > 0, and a comment on why the result should be trusted despite the formal derivation. With that, it's publishable.","headline":"Explicit S/T matrices for Carrollian and warped CFTs, but the derivation's key step is only justified on the real axis.","tokens_in":18055,"tokens_out":4117,"would_cite":false,"duration_ms":41568,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["11.25.Hf"],"model":"deepseek-v4-flash","headline":"The paper derives explicit modular S and T matrices for 2d Carrollian and warped CFT characters, showing S^2=1 for Carrollian theories and S^2 equal to charge conjugation for warped theories.","keywords":["modular S-matrix","modular T-matrix","Carrollian CFT","warped CFT","bms3 algebra","Virasoro u(1) Kac-Moody","character transformations","charge conjugation"],"falsifier":"Take the highest-weight Carrollian S matrix (4.16), insert it into the defining equation (4.7) with tau = i (so Im tau > 0), and numerically evaluate the double integral; if the result differs from the character at (-1/i, rho/$i^{2}$) by more than numerical error, the claimed S matrix fails outside the real-tau line where the derivation was performed.","tokens_in":16975,"feed_emoji":"🔁","tokens_out":7663,"duration_ms":61338,"temperature":0.7,"pith_summary":"This paper tries to establish explicit modular S and T matrices for the characters of two families of non-relativistic two-dimensional conformal field theories: Carrollian CFTs, whose symmetry is the bms3 algebra, and warped CFTs, whose symmetry is one Virasoro copy in a semi-direct product with a u(1) Kac-Moody algebra. The derivations cover several representation families: highest-weight and induced for Carrollian, and non-unitary (imaginary charge) and unitary for warped. The S matrices are written as explicit integral kernels in Liouville-like momenta, while the T matrices are phase factors times delta functions. A central check is that these kernels implement the character transformations, with $S^{2}$=1 for Carrollian characters and, for warped characters, SS^dagger=1 and $S^{2}$=C where C flips the u(1) charge. The motivation is that these theories are candidate holographic duals for flat and warped three-dimensional gravity, where modular data controls Cardy-like growth and near-extremal behaviour.","feed_headline":"Modular S, T matrices derived for 2d Carrollian and warped CFTs","feed_subtitle":"Explicit kernels give the modular S and T action on Carrollian and warped characters, with charge conjugation for warped.","key_machinery":"The engine is a delta-function localization: for characters of the form $\\chi=\\exp[2\\pi i\\tau P_\\tau]\\exp[2\\pi i z s(p)]/G(\\eta(\\tau))$, an ansatz $S=A(p,p')\\exp[-2\\pi i(\\gamma(p,p')P_\\tau+\\gamma(p',p)P'_\\tau)]$ produces a delta function $\\delta(\\tau-\\gamma(p,p'))$ that localizes the $p$-integral. Matching the remaining $P'_\\tau$ and $z$ exponentials fixes $\\gamma(p',p_\\star)=1/\\tau$ and $s(p_\\star)/s(p')=z'/z$, and the normalization is fixed by the modular transformation of $G(\\eta(\\tau))$, typically through the eta-function identity $\\eta(-1/\\tau)=\\sqrt{-i\\tau}\\eta(\\tau)$. This mechanism turns the modular S-transformation into a Fourier-like kernel without assuming factorization of the characters.","core_discovery":"The central discovery is that the characters of 2d Carrollian and warped CFT algebras admit modular S and T matrices of a definite, closed form, provided the quantum numbers are parametrized by Liouville-type momenta. For Carrollian highest-weight representations the non-vacuum S kernel is $S_{\\rm car}(P'_L,P'_M;P_L,P_M)=2\\frac{|P'_M|}{|P_M|^2}\\frac{P_M}{P'_M}\\sin\\left[2\\pi\\left(\\frac{P'_M}{P_M}P_L+\\frac{P_M}{P'_M}P'_L\\right)\\right]$, the induced-representation kernel is $S_{\\rm car}^{\\rm ind}=2/|P_M|\\cos[2\\pi(\\frac{P'_M}{P_M}P_L+\\frac{P_M}{P'_M}P'_L)]$, and the warped kernels, for imaginary and real charge, are $S=i\\frac{p}{p'}\\frac{|p'|}{|p|^2}\\exp[-2\\pi i(\\frac{p'}{p}P_W+\\frac{p}{p'}P'_W)]$. The paper verifies that these kernels transform the characters correctly under $\\tau\\to-1/\\tau$ (with the appropriate transformation of the second modular parameter) and that $S^2=1$ for the Carrollian kernels, while for the warped kernels $S S^\\dagger=1$ and $S^2$ equals charge conjugation, $p\\to-p$. The T matrices are computed as $e^{-i\\pi/6}\\delta(\\ldots)$ or $e^{-i\\pi/12}\\delta(\\ldots)$ for the respective cases. The derivation does not rely on character factorization, and the paper notes that the method breaks down for non-unitary warped CFTs with real charge because the descendant factor $\\eta(2\\tau)$ lacks a universal modular transformation.","pith_inferences":["If the S matrices hold beyond the real-tau line, the continuous-spectrum Fourier-like kernels suggest that any fusion or braiding coefficients for these algebras, if they exist, will also take the form of oscillatory integral kernels, a natural next step the paper does not pursue.","The failure for non-unitary real-charge warped CFTs (due to eta(2tau)) hints that a full modular theory for that sector may require generalized eta functions or a different parameterization; one testable extension is to look for a kernel using eta(tau/2) or a Jacobi-form completion.","Because the method produces S^2=C for warped theories, one can predict that modular invariant partition functions in these theories, when combined with a charge-conjugation invariant spectrum, must project onto the C-invariant sector; this could be checked against known warped CFT partition functions.","The delta-function localization is formally a stationary-phase/plane-wave decomposition, so the same technique might apply to other solvable characters (e.g., higher-spin or BMS-type algebras) provided the descendant factor G(eta) has a known modular transformation."],"forward_implications":["The Carrollian modular S matrices (4.16) and (4.32) satisfy S^2=1 when acting on characters, reproducing the group relation of the Carrollian modular group.","The warped modular S matrices (5.12) and (5.16) satisfy SS^dagger=1 and S^2=C, where C implements p to -p, showing the characters form a non-faithful (projective) representation of the modular group due to the u(1) symmetry.","The T matrices are trivial up to a phase: e^{-i pi/6} or e^{-i pi/12} times delta functions in the quantum numbers, so the characters' T-transformation comes entirely from the Dedekind eta factor.","For Carrollian induced representations, the absolute value |eta(sigma)| makes characters T-invariant, so the T matrix is just the identity kernel.","The S kernels provide a route to the density of primary states (Cardy-like growth) in these theories, as the paper states it will report elsewhere."],"supporting_citations":[{"why":"Supplies the highest-weight Carrollian character formula (4.5) that the Carrollian S-matrix derivation starts from.","marker":"[43]"},{"why":"Supplies the induced-representation Carrollian character formula (4.29) whose S and T matrices are derived.","marker":"[48]"},{"why":"Supplies the non-unitary and unitary warped characters (5.5), (5.7), (5.15) used to fix the warped S and T kernels.","marker":"[49]"},{"why":"Supplies the notation and near-extremal context for warped characters, including the real/imaginary charge split.","marker":"[45]"},{"why":"Defines the warped CFT modular group action and the weak Jacobi form transformation used for the S-transformation in (5.2).","marker":"[9]"},{"why":"Provides the Carrollian modular group action (4.2) and the torus modular parameters (sigma, rho).","marker":"[13]"},{"why":"Supplies the Gaussian-integral identity underlying the Fourier-kernel derivation of the Virasoro S matrix, which the paper's strategy generalizes.","marker":"[50]"},{"why":"Source for the defining properties of modular S matrices (S^2=1 and the charge-conjugation modification S^2=C in the presence of extra symmetries).","marker":"[3]"}],"fun_headline_variants":["Explicit modular S, T for Carrollian and warped CFTs","Liouville momenta unlock modular kernels for warped CFTs","Closed-form S and T matrices for Carrollian characters","Modular S, T with charge conjugation for warped CFTs"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The derivation assumes the delta-function localization, which is only rigorously valid for real tau, can be extended to the upper half-plane where torus modular transformations are defined; the paper gives no analytic-continuation argument.","fun_headline_variants_meta":{"raw":{"variants":["Explicit modular S, T for Carrollian and warped CFTs","Liouville momenta unlock modular kernels for warped CFTs","Closed-form S and T matrices for Carrollian characters","Modular S, T with charge conjugation for warped CFTs"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000175,"raw_usage":{"total_tokens":1291,"prompt_tokens":956,"completion_tokens":335,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":572,"completion_tokens_details":{"reasoning_tokens":260}},"tokens_in":572,"tokens_out":335,"duration_ms":3581,"temperature":1.0,"reasoning_tokens":260,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T17:11:48.666933+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take the highest-weight Carrollian S matrix (4.16), insert it into the defining equation (4.7) with tau = i (so Im tau > 0), and numerically evaluate the double integral; if the result differs from the character at (-1/i, rho/$i^{2}$) by more than numerical error, the claimed S matrix fails outside the real-tau line where the derivation was performed.","supporting_citations":[],"review_version":1}