{"id":"95275bba-eda8-497a-92de-71e7931aab13","arxiv_id":"2608.05285","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"A CSW-like recursion plus a 'contact Lagrangian' computes arbitrary tree-level N-photon amplitudes in general EFTs, with results through 10 photons in Born-Infeld theory.","lead":"This paper builds a new recursive method for computing photon scattering amplitudes in any effective field theory of electromagnetism, using self-dual fields and local 'contact' building blocks. It produces compact formulas up to 10 photons and reveals a simple link between electromagnetic duality and helicity conservation.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The unconstrained-field reformulation (Sec. 6.1.1) underlying the contact Lagrangian is asserted rather than proven; if the F=f+φ split misses Bianchi constraints or introduces spurious modes, all sewn higher-point amplitudes inherit the error.","rationale":"The central claim—that all N-photon tree amplitudes are determined by the contact Lagrangian—rests entirely on the unconstrained reformulation of Sec. 6.1.1. Every higher-point result (contact Lagrangian up to N=14, 8-pt generic amplitudes, 10-pt BI amplitudes) is obtained by integrating out φ± and then sewing the resulting f± vertices. If the claimed equivalence between the constrained integral (6.6) and the unconstrained one (6.7) is not exact, the error propagates into all these results. The paper's own evidence for the reformulation consists of the 6-pt consistency checks in App. C and agreement of the BI contact Lagrangian with [76]; these are real but do not exercise the region where the reformulation is most dangerous: 8-point contact amplitudes with non-conserving helicity (e.g., A_c[6+,2−]) that in a helicity-conserving theory would need to cancel against sewn diagrams. The alternative candidate concern—the 'only if' direction of the duality/helicity equivalence (Sec. 7.1)—is a logical gap in a conceptual claim, but it does not threaten the computational results themselves, since the BI bootstrap uses duality invariance rather than the converse. Thus the reformulation is the single load-bearing assumption. The proposed direct Feynman-diagram computation at 8 points would settle it. My assessment agrees with the reader's weakest assumption, so the CONDITIONAL verdict stands unless the test is performed.","tokens_in":48733,"tokens_out":16433,"duration_ms":156564,"concrete_test":"Compute the 8-point contact amplitudes A_c[6+,2−] and A_c[4+,4−] directly from the original EFT Lagrangian (2.2) by summing all tree diagrams that contain only ++/-- propagators, using standard Feynman rules with the κ coefficients. Compare the results with the Hafnian formula (4.22) using the contact coefficients α̂(8)_6 and α̂(8)_4 from (6.22) and App. E. If they disagree for generic momenta and generic κ, the unconstrained reformulation is not faithful and the derived amplitudes in Sec. 8 are invalid. If they agree, the reformulation is validated at the first nontrivial order where it is load-bearing.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central machinery depends on the claim (Sec. 6.1.1) that constrained self-dual fields F± can be replaced by unconstrained pairs f±, φ± with F±=f±+φ±, propagator split (6.2), and vanishing mixed correlators ⟨fφ⟩. The equivalence to the original path integral (6.6) is asserted via (6.7) without specifying the measure, the reality properties of f± and φ±, or the Jacobian. If f and φ are genuinely independent complex fields, the path integral has twice the degrees of freedom of the real EFT; if they are constrained by φ−=(φ+)∗, the 'unconstrained' claim fails and the kinetic term (6.3) is not real. This is not a mere technicality: the contact Lagrangian (6.22), the all-order resummation (7.17), and the 8-/10-point amplitudes in Sec. 8 are all derived from this reformulation. The paper's own 6-point checks (App. C) and agreement with [76] are encouraging but do not probe the first multiplicity (8 points) at which non-conserving contact vertices coexist with sewn topologies and could mask a mismatch. A direct Feynman-diagram comparison at 8 points would settle whether the reformulation is faithful.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops a recursive (CSW-like) method for tree-level N-photon helicity amplitudes in the general low-energy EFT of electromagnetism. It works in the self-dual basis, shows that the ±± field-strength propagators are contact terms while the +− propagator carries the pole, and uses the resulting diagrammatic substructure to decompose any amplitude into a web of local contact amplitudes joined by +− lines. Contact amplitudes are encoded in a weighted Hafnian formula (Eq. (4.22)) and in a 'contact Lagrangian' obtained by decomposing F± = f± + φ± and integrating out the non-propagating φ±. The paper derives the contact Lagrangian to O(F^14), proves (within its formalism) equivalence between off-shell electromagnetic duality and helicity conservation, resums the BI contact Lagrangian in closed form, and presents complete 6-pt and 8-pt generic EFT amplitudes and 10-pt BI amplitudes.","tokens_in":49054,"tokens_out":8691,"duration_ms":80159,"significance":"Should the unconstrained reformulation be valid, this is a substantial technical advance: it turns a combinatorial Feynman-diagram problem into a small set of Hafnians, provides compact closed-form BI results, and gives a simple proof of the duality/helicity-conservation equivalence. The paper includes explicit Feynman-rule cross-checks at 6 points (App. C) and agreement with the existing literature [70,76], which are genuine strengths. The central risk is the unproven path-integral equivalence in Sec. 6.1.1; because all higher-point results inherit from it, this needs to be established before the completeness claims can be accepted.","major_comments":[{"comment":"The central equivalence between the Bianchi-constrained path integral (6.6) and the unconstrained f,φ path integral (6.7) is asserted rather than proven. The paper does not specify the integration measure over constrained F±, the reality properties of f± and φ±, or the Jacobian of the change of variables; if f and φ are independent complex fields, the field content is doubled relative to the real EFT, while if they are related by reality conditions the 'unconstrained' characterization is inaccurate. Since the contact Lagrangian (6.22), the BI resummation (7.17), and the 8- and 10-pt amplitudes of Sec. 8 all follow from this reformulation, a proof of equivalence—or at least a direct 8-pt Feynman-diagram comparison, where contact and sewn topologies first coexist—is required. The 6-pt checks in App. C are encouraging but do not settle this point.","section":"Sec. 6.1.1, Eqs. (6.1)–(6.7)"},{"comment":"The advertised 'complete' 8-pt and 10-pt amplitudes are written as topology sums in which the 'sum over channels' is not expanded, and the weighted Hafnian cHf involving off-shell internal momenta q is not explicitly defined (the off-shell spinor continuation is only sketched around Eq. (5.1)). As a result, no reader can verify that the channel sums are exhaustive or reproduce the displayed expressions. The authors should either expand the channel sums explicitly or provide a machine-readable ancillary file with the full expansions, together with the off-shell definition of cHf.","section":"Secs. 8.1 and 8.2, Eqs. (8.6)–(8.12)"},{"comment":"The bootstrap of the BI contact Lagrangian leaves a free parameter a, which is then set to a=1 to match Refs. [76,80]. No physical condition (e.g., reality, analyticity, or the absence of spurious singularities) is shown to select a=1, so the derivation is not self-contained: the final closed form (7.17) is in effect fitted to known results. The authors should either derive the value of a from a stated condition or explicitly present the a-family as an ansatz whose agreement with BI is verified a posteriori.","section":"Sec. 7.3, Eqs. (7.8)–(7.10)"},{"comment":"The claim that all loop contributions to the contact Lagrangian vanish 'identically upon using dimensional regularization' is too terse to support the all-loop statements in Secs. 6.1.4 and 7.1. Loop integrals with only local propagators are polynomials in loop momenta, but the argument should spell out the scaleless-integral structure and clarify that f-loop contributions to scattering amplitudes are not being discarded.","section":"Sec. 6.1.3, Eq. (6.11)"}],"minor_comments":[{"comment":"The phrase 'A a CSW-like recursion' contains a typo and should read 'A CSW-like recursion'.","section":"Introduction, p. 4"},{"comment":"The displayed constraint appears garbled: binary operators are missing and there is a stray '−h−' in the second term; please re-typeset and re-check the equation.","section":"Eq. (7.8)"},{"comment":"The statement that the source is chosen to have 'no constant component' should be made precise (e.g., J(q=0)=0 in momentum space), and its compatibility with the LSZ procedure should be stated.","section":"Sec. 6.1.3"},{"comment":"The O(F^14) contact Lagrangian is listed without derivation; given its length, a generating script or ancillary file would aid reproducibility and reduce the risk of typographical errors in the coefficients.","section":"App. E"},{"comment":"The notation [i|q|j⟩ should be accompanied by an explicit definition of the all-incoming momentum routing for the internal line q, including the sign convention used in the sewing formula.","section":"Eq. (5.1) and Sec. 5.1.2"}],"recommendation":"major_revision","confidential_remarks":"The paper's relationship to Ref. [76] is very close: the contact Lagrangian, the static-field integration, the duality variables z±, and the closed-form BI resummation all appear there. The present work's added value lies in the CSW-style sewing framework, the Hafnian representation, and the 8/10-pt results. I recommend that the editors ask the authors to sharpen the novelty statement relative to [76] and to make the 8-pt channel sums available in a verifiable form."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Worth a careful look. The paper gives a recursive construction of tree photon amplitudes in the general EFT of electromagnetism, built from local contact amplitudes written as weighted Hafnians, and provides the contact Lagrangian up to N=14. The Born-Infeld resummation via Lagrange inversion is a clean piece of work and agrees with [76]. The 6-point forward-limit checks pass, and the central idea — using the self-dual basis to isolate contact vertices as the only non-factorizable building blocks — is genuinely new. It is a real step beyond the usual soft-bootstrap approach to EFT amplitudes.\n\nThe main soft spot is the unconstrained-field reformulation in Sec. 6.1.1. The paper splits F± = f± + φ± and asserts that mixed correlators vanish and that this reproduces the Bianchi-constrained path integral. That is the load-bearing step for the contact Lagrangian, the all-order resummation, and the 8-/10-point results. It is stated more than proven: the measure and reality properties of f and φ are not specified, and the Jacobian of the change of variables is not addressed. The 6-point checks are encouraging but they do not probe the 8-point level where sewn topologies and non-conserving contact vertices coexist; that is exactly where a mismatch would show up. So the central claim is plausible, not yet proven.\n\nThe 8-point and 10-point results are written as sums over channels that are not expanded, which makes independent verification hard. The all-loop tree-exactness argument is terse; the claim that loop graphs vanish in dimensional regularization because the φ propagators contain no denominators is reasonable, but a few explicit examples would help. The citation of [76] is honest, and the bootstrap parameter a=1 is a choice rather than a fitted crutch, so circularity is not a concern.\n\nNone of this is a deal-breaker. The paper is self-contained, the agreements with [76] and the 6-point Feynman-rule checks give real evidence, and the explicit contact Lagrangian in App. E is a useful resource. It deserves a serious referee. I would send it to review, asking the referee to press on the unconstrained reformulation — either a fuller derivation or an explicit 8-point check in a tractable kinematic limit. I would also cite the BI contact Lagrangian and the Hafnian representation in my own work.","headline":"A serious new method for photon EFT amplitudes with a plausible but under-proven unconstrained-field reformulation.","tokens_in":49547,"tokens_out":4543,"would_cite":true,"duration_ms":39955,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["81T18","81T13","81U20"],"pacs":["11.55.-m","12.20.-m"],"model":"deepseek-v4-flash","headline":"This paper establishes a recursion that builds every tree-level photon amplitude in any EFT of electromagnetism from local contact amplitudes, each a single weighted Hafnian.","keywords":["photon amplitudes","effective field theory","self-dual basis","helicity amplitudes","Hafnian","Born-Infeld electrodynamics","electromagnetic duality","CSW recursion"],"falsifier":"Compute a specific 6-point helicity amplitude, say $A[5^+,1^-]$, in the generic EFT directly from Feynman rules and compare it order by order with the sewn Hafnian channel sum (5.9); any mismatch in the channel sum would show that the contact decomposition is incomplete.","tokens_in":115,"feed_emoji":"⚛️","tokens_out":6935,"duration_ms":121399,"temperature":0.7,"pith_summary":"This paper tries to establish that every tree-level amplitude with $N$ photons in any effective field theory of electromagnetism is built from purely local contact amplitudes joined by mixed-helicity photon propagators, with each contact amplitude equal to a single weighted Hafnian. The authors show that these contact amplitudes are encoded in a contact Lagrangian computed by integrating out the non-propagating self-dual components of the field strength. If the construction is right, it gives a practical recursion for photon amplitudes that bypasses Feynman-diagram combinatorics, and it exposes why helicity conservation and off-shell electromagnetic duality coincide in these theories. Explicit outputs include contact Lagrangians up to 14 photons, complete 6- and 8-point amplitudes for the generic EFT, and Born-Infeld amplitudes up to 10 points.","feed_headline":"All photon amplitudes in any EFT are webs of single Hafnians","feed_subtitle":"A self-dual recursion plus a contact Lagrangian computes photon scattering up to 10 points in Born-Infeld.","key_machinery":"The load-bearing mechanism is the decomposition $F_\\pm = f_\\pm + \\phi_\\pm$ of the field strength together with the resulting contact Lagrangian $L_c[f_\\pm]$. The $f_\\pm$ fields carry the propagating mixed-helicity lines, while the $\\phi_\\pm$ fields have purely local propagators and are integrated out; the operators of $L_c$ are in one-to-one correspondence with the contact amplitudes, each of which evaluates to a weighted Hafnian because of the pairing structure of the EFT vertices.","core_discovery":"In the self-dual basis the photon only propagates through the mixed $+$/$-$ channel; the same-helicity propagators are local, so every amplitude factorizes into a web of local contact subamplitudes connected by $+$/$-$ lines. A contact subamplitude with $K$ positive and $N-K$ negative helicities is the weighted Hafnian $\\widehat{\\mathrm{Hf}}\\big(\\chi^{(N)}_K\\big)$ of the angle/square bracket matrix, with the weight set by photon multiplicities. The contact Lagrangian obtained by integrating out the static $\\phi^\\pm$ fields contains one operator for each contact amplitude, and its coefficients are produced by a fixed-point iteration with a known stopping depth. For helicity-conserving theories the contact Lagrangian must be built from $(f_+^2 f_-^2)^P$ terms, which is exactly the statement of invariance under $f_\\pm \\to e^{\\mp i\\theta} f_\\pm$; this yields the claimed equivalence between off-shell electromagnetic duality and helicity conservation. In Born-Infeld theory the contact Lagrangian resums in closed form by Lagrange inversion.","pith_inferences":["The same decomposition might be adapted to non-abelian gauge theories or gravity by replacing the scalar contact Lagrangian with a matrix-valued one, though the paper does not pursue that extension.","Because the weighted Hafnian is a permanent-like sum over perfect matchings, the computational cost of generating contact amplitudes is combinatorial; for very large $N$ the recursion could require approximation or tensor-network methods, a limitation the paper does not discuss.","The equivalence between duality and helicity conservation suggests that any UV completion preserving off-shell duality at the amplitude level must have a contact Lagrangian of the $(f_+^2 f_-^2)^P$ form, which could be tested by computing helicity-violating amplitudes in candidate completions."],"forward_implications":["All 8-point tree amplitudes of the generic EFT of electromagnetism are determined by the displayed sums of Hafnian products over channels, with no Feynman-diagram bookkeeping.","The 10-point Born-Infeld amplitude $A[5^+,5^-]$ is computed from 4-point contact blocks alone, since helicity conservation forbids all other 10-point amplitudes.","For a fixed number of external photons only a finite, known number of contact amplitudes is needed; the stopping criterion $m_{\\min} = \\lfloor (N-2)/4 \\rfloor$ bounds the required iteration depth.","Helicity conservation and off-shell electromagnetic duality are equivalent for the general EFT, both at tree and loop level, because the contact Lagrangian takes the duality-invariant form if and only if no helicity-violating contact operator appears.","The contact Lagrangian receives only tree-level contributions, so the recursion framework is not limited to tree order and can be used to organize higher-loop correlators."],"supporting_citations":[{"why":"Supplies the CSW-style recursion pattern that the sewing of contact amplitudes mimics, providing the factorization logic.","marker":"[1]"},{"why":"Supplies the self-dual field-to-helicity mapping and the low-energy Euler-Heisenberg amplitudes used for consistency checks.","marker":"[70]"},{"why":"Supplies the tree-level helicity conservation of Born-Infeld from duality invariance, which the contact-Lagrangian proof reformulates.","marker":"[78]"},{"why":"Supplies the earlier normal-ordering and contact-Lagrangian derivation of helicity conservation in nonlinear QED that the present functional method reproduces and extends.","marker":"[76]"},{"why":"Supplies the general EFT-of-electromagnetism setup and the soft-bootstrap perspective in which the vector recursion is framed.","marker":"[42]"}],"fun_headline_variants":["EFT photon amplitudes as Hafnian webs from self-dual recursion","All EFT photon amplitudes factorization via Hafnian recursion","Hafnian contact terms power recursive EFT photon amplitudes","Born-Infeld photon amplitudes resummed via contact Lagrangian"],"cache_read_input_tokens":51584,"weakest_assumption_plain":"The construction assumes that replacing the Bianchi-constrained fields by unconstrained $f_\\pm,\\phi_\\pm$ and integrating out $\\phi_\\pm$ gives a contact Lagrangian that reproduces the original EFT exactly; if that equivalence fails, every sewn amplitude would miss the physics of the original theory.","fun_headline_variants_meta":{"raw":{"variants":["EFT photon amplitudes as Hafnian webs from self-dual recursion","All EFT photon amplitudes factorization via Hafnian recursion","Hafnian contact terms power recursive EFT photon amplitudes","Born-Infeld photon amplitudes resummed via contact Lagrangian"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000668,"raw_usage":{"total_tokens":3077,"prompt_tokens":1006,"completion_tokens":2071,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":622,"completion_tokens_details":{"reasoning_tokens":2012}},"tokens_in":622,"tokens_out":2071,"duration_ms":14123,"temperature":1.0,"reasoning_tokens":2012,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-08T16:19:19.593291+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute a specific 6-point helicity amplitude, say $A[5^+,1^-]$, in the generic EFT directly from Feynman rules and compare it order by order with the sewn Hafnian channel sum (5.9); any mismatch in the channel sum would show that the contact decomposition is incomplete.","supporting_citations":[],"review_version":1}