{"id":"8d813ba5-e919-4b62-86af-929c325b8b4d","arxiv_id":"2507.01600","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For massive Klein-Gordon operators with asymptotically static potentials on asymptotically Minkowski spacetimes, the authors construct a unique Feynman propagator and prove a microlocal Hadamard wavefront condition.","lead":"This mathematics paper constructs the Feynman propagator, a special inverse of the massive Klein-Gordon equation, on spacetimes that become flat Minkowski space at infinity but carry static or static-like potentials. The construction uses Vasy's 3sc microlocal calculus, avoids global time-splitting assumptions, and proves the desired wavefront regularity.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 6.3 proves only a one-sided wavefront inclusion, while Definition 6.1 and the abstract promise the kernel-level Radzikowski equality; the gap is unaddressed and load-bearing for the advertised microlocal Hadamard condition.","rationale":"The reader's weakest_assumption is the dynamical structure theorem and global source-sink flow. I do not find the imported propagation estimates or the flow-geometry assumption to be the most exposed point, since they are stated as theorems from the companion paper and the non-trapping hypothesis is explicit. The most load-bearing gap is internal: the paper defines the microlocal Hadamard condition as a kernel-level equality (Definition 6.1) but proves only a solution-wise one-sided wavefront inclusion (Theorem 6.3). This gap is not merely expository: the domain spaces impose above-threshold regularity at Rsrc, and no argument shows that this restriction is invisible at the level of the full Schwartz kernel. The claim is therefore stronger than what is established. This is exactly the kind of missing-support issue that warrants a conditional verdict rather than acceptance. Because the reader already assigned CONDITIONAL and flagged this concern as item (1), though not as the weakest assumption, my read does not move the verdict. The proposed free-case computation would settle whether the equality actually holds in the model case and would isolate the remaining proof obligation for general V.","tokens_in":31678,"tokens_out":23959,"duration_ms":277130,"concrete_test":"For V=0 on Minkowski space, compute explicitly the Schwartz kernel of the inverse (P0)_Fey^{-1} produced by the Section 2.4 Fredholm construction (or its 3sc analogue) and compare its wavefront set ~WF((P0)_Fey^{-1}) with diag ∪ ⋃_{s≥0} Φ^s(diag Char(P0)). If the kernel-level equality fails, the central Hadamard claim is false even in the free case. If it holds, the free case supports the equality; one then still needs a kernel-level propagation argument to extend the equality to admissible V. A minimal sub-check is to test the diagonal component over Rsrc: apply the inverse to a compactly supported distribution whose wavefront set approaches Rsrc and confirm whether the diagonal singularity of the kernel is present there.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The manuscript defines a Feynman parametrix in Definition 6.1 by the kernel-level equality ~WF(E) = diag ∪ ⋃_{s≥0} Φ^s(diag Char(P0)), and the abstract and introduction advertise that the constructed (PV)_Fey^{-1} satisfies this 'microlocal Hadamard condition'. However, the only statement proved, Theorem 6.3, is the one-sided inclusion WFcl((PV)_Fey^{-1}f) ⊂ WFcl(f) ∪ ⋃_{s≥0} Φ^s(WFcl(f)∩Char(P0)) for solutions with compactly supported forcing, and its proof is one sentence: 'This theorem follows directly from the propagation estimates in the scattering calculus.' Proposition 6.2 shows the inclusion is a consequence of the kernel equality, but the converse is not established: Hörmander's composition theorem gives only the containment ~WF(E) ⊂ diag ∪ ⋃_{s≥0} Φ^s(diag Char(P0)) from the solution-wise inclusion; it gives no lower bound. The paper never shows that every point of diag, and every forward-flowed characteristic pair, actually belongs to ~WF(E). If any such point is absent—for example, a diagonal point over the radial sources, where the domain spaces impose above-threshold regularity—the constructed inverse would not be a Feynman parametrix in the paper's own sense, and the advertised Hadamard condition would fail. Thus the central claim as stated overreaches its proof.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper constructs Feynman propagators for massive Klein-Gordon operators with asymptotically static potentials on asymptotically Minkowski spacetimes. The authors set up a Fredholm problem PV: X_{s,ℓ0,ℓ+} → Y_{s,ℓ0,ℓ+} in the 3sc-calculus, using microlocal cutoffs at the radial sources of the Hamiltonian flow in place of variable-order Sobolev spaces. They prove Fredholmness and, under self-adjointness and positivity or absence of bound states of the limiting spatial Hamiltonians, invertibility; the inverse is shown to be independent of the parameters and cutoffs. They also prove a one-sided classical wavefront set inclusion for solutions with compactly supported forcing, and an analogous bound-state result. The claimed novelty is a global Feynman propagator for asymptotically static potentials without a global time-splitting or time-slice spectral assumptions.","tokens_in":31939,"tokens_out":11096,"duration_ms":127374,"significance":"The Fredholm framework and the avoidance of variable-order weights by microlocal cutoffs are substantial technical contributions. If the results are correct, the paper would provide the first construction of Feynman propagators at this level of generality for asymptotically static potentials, extending work of Gérard–Wrochna and Vasy. The paper is careful about independence of cutoffs and about bound-state obstructions. However, the advertised microlocal Hadamard condition is only proved in a one-sided form; the kernel-level equality in Definition 6.1 is not established, so the paper's central claim as stated is stronger than the proof.","major_comments":[{"comment":"Definition 6.1 defines a Feynman parametrix by the kernel-level equality ~WF(E) = diag(T*X\\0) ∪ ⋃_{s≥0} Φ^s(diag Char(P0)). Proposition 6.2 derives from this equality only the one-sided inclusion WFcl(Ef) ⊂ WFcl(f) ∪ ⋃_{s≥0} Φ^s(WFcl(f)∩Char(P0)), and Theorem 6.3 proves exactly that one-sided inclusion, with the proof given as a single sentence: 'This theorem follows directly from the propagation estimates in the scattering calculus.' No argument is supplied for the reverse inclusion: namely, that every point of the diagonal and every forward-flowed characteristic pair actually belongs to ~WF((PV)_Fey^{-1}). Hörmander's composition theorem gives only the containment direction from a solution-wise inclusion, not equality. Since the abstract and introduction advertise the microlocal Hadamard condition, this gap is load-bearing. The authors should either prove the equality at the kernel level, or explicitly weaken the claim and Definition 6.1 to a one-sided singularity-propagation statement.","section":"Section 6, Definition 6.1 and Theorem 6.3"},{"comment":"The global Fredholm estimate in Lemma 5.2 is the technical core of Theorem 5.1, but its proof is not fully self-contained. In particular, the proof uses an open cover O1,...,O4 of the compressed cotangent bundle with six listed properties, and the existence of this cover is asserted with 'The proof that such a cover exists is similar to the case in [2].' Since the paper deliberately departs from the variable-order setting of [2], it should either state this cover construction as a precise lemma with the hypotheses used here, or give a detailed proof that the estimates of Propositions 4.8–4.10 combine in exactly this fixed-weight setting. As written, the central Fredholm estimate rests on an unstated technical lemma.","section":"Section 5, Lemma 5.2 and Theorem 5.1"}],"minor_comments":[{"comment":"The sentence 'Near NP we allow asymptotics as in (40) with τ0 > 0 and exclude those with τ0 < 0, while near NP we exclude those with τ0 < 0' contains a typo: the second 'NP' should be 'SP'.","section":"Section 7, paragraph after (40)"},{"comment":"Theorem 2.3 is stated without proof in Section 2; the authors should either prove the free-case invertibility directly or explicitly refer to the general proof in Section 5, since the free case is covered by the later argument.","section":"Section 2, Theorem 2.3"},{"comment":"The notation for wavefront sets is inconsistent: Theorem 2.3 uses WF, while Definition 6.1 and Theorem 6.3 use WFcl. Please unify the notation and define both classical and 3sc-wavefront sets clearly in one place.","section":"Throughout, Theorem 2.3 and Theorem 6.3"},{"comment":"The sentence 'The inverse (PV)_Fey^{-1} as constructed in Proposition 5.7 has the same property' is ambiguous, because Proposition 6.2 states a one-sided inclusion, whereas Definition 6.1 states an equality. The authors should specify that the inverse satisfies the one-sided inclusion only.","section":"Section 6, after Proposition 6.2"},{"comment":"The proof of Lemma 5.2 would be easier to verify if the interpolation argument and the absorption of the ∥QsrcGϕu∥_{s−1,ℓ′} term were written out rather than referenced to [2, Eq. (2.39)], especially because the spaces here are not the variable-order spaces of [2].","section":"Section 5, Lemma 5.2"}],"recommendation":"major_revision","confidential_remarks":"The main issue is the mismatch between Definition 6.1 and Theorem 6.3: the paper advertises the microlocal Hadamard condition but proves only a one-sided wavefront inclusion. This is fixable by either adding a kernel-level proof of equality or by carefully restating the main claim as a one-sided propagation result. The Fredholm construction itself appears sound and is a substantial contribution, so I would not recommend rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nThe headline: the Fredholm construction is real, and this is the first treatment of asymptotically static potentials without a global time-splitting, but the advertised microlocal Hadamard condition overreaches the proof. Theorem 6.3 gives a one-sided wavefront inclusion for Feynman solutions; Definition 6.1 defines a Feynman parametrix by a kernel-level equality. The paper never proves that equality, and Hörmander's composition theorem will not supply the lower bound. The abstract and introduction say \"a microlocal Hadamard condition,\" which is arguably weaker, but the formal definition and the theorem do not match. That is a gap the authors need to close or explicitly disclaim.\n\nWhat is genuinely good: replacing variable order Sobolev spaces with microlocal cutoffs at the radial sources is a clean idea, and it does sidestep the incompatibility between variable weights and the indicial family. The free case in Section 2 is a useful exposition. The Fredholm setup, once you accept the estimates imported from [2], is coherent: kernel and cokernel finiteness follow from the global estimates, and invertibility under positivity of the limiting spatial Hamiltonians is argued in detail. The self-citations to [2] are legitimate; these are prior results, not assumptions containing the present theorem.\n\nSoft spots: (1) the wavefront gap above, which is the most serious; (2) the proof leans heavily on Propositions 4.8–4.10 from [2], and readers of this paper will need to go back to that paper to verify; that is normal for a sequel, but the paper could summarize the key estimates more; (3) the existence of the microlocal partition of unity cover in Lemma 5.2 is asserted in half a paragraph citing [2, p. 83]—I would like to see the argument sketched, since the cover has to respect the flow control conditions. These are addressable.\n\nWho this is for: microlocal analysts and mathematical physicists working on QFT on curved spacetime. The reader who wants to know how to get a Feynman propagator for static-ish potentials without spectral assumptions per slice will get real value. A serious referee should get this paper. My recommendation: send it to a good journal, but ask the authors to reconcile the Hadamard statement with the one-sided inclusion, or to rephrase Definition 6.1 to match what is actually proved.","headline":"The Fredholm construction is a genuine step forward, but the paper overclaims the Hadamard condition: Theorem 6.3 proves only a one-sided wavefront inclusion, while Definition 6.1 promises a kernel-level equality.","tokens_in":32529,"tokens_out":3025,"would_cite":true,"duration_ms":35125,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35L05","35Q40","58J40","81T20"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper constructs a global Feynman propagator for the massive Klein-Gordon operator with asymptotically static potentials on asymptotically Minkowski spacetimes, and proves it satisfies a microlocal Hadamard condition.","keywords":["Feynman propagator","Klein-Gordon equation","asymptotically static potential","microlocal Hadamard condition","3sc-calculus","propagation of singularities","radial points","Fredholm theory"],"falsifier":"Check a concrete asymptotically Minkowski metric whose compactified Hamiltonian flow admits a trapped bicharacteristic; if the global Fredholm estimate of Lemma 5.2 still holds, the framework is more general, whereas if it fails, this paper's Feynman construction does not cover that spacetime. Also compute the indicial family $\\tau^2 - H_{V_\\pm}$: if $0$ is an eigenvalue of either limiting Hamiltonian, the main invertibility theorems do not apply.","tokens_in":31437,"feed_emoji":"⚛️","tokens_out":5183,"duration_ms":58669,"temperature":0.7,"pith_summary":"The paper aims to show that the Klein-Gordon operator $P_V = \\square_g - m^2 - V$, where $V$ is a decaying potential that becomes static at timelike infinity, admits a well-defined global Feynman propagator on asymptotically Minkowski spacetimes. It claims that $P_V$ is a Fredholm map between specially designed Sobolev-type spaces, and that under natural hypotheses (self-adjointness and positivity of the limiting spatial Hamiltonians) this map is invertible. The inverse is shown to be independent of all parameters and cutoffs used in its construction, and to satisfy the wavefront-set inclusion that characterizes the Feynman propagator. If correct, this gives a robust route to quantum-field-theoretic propagators on curved backgrounds without imposing global spectral hypotheses on each time slice.","feed_headline":"Feynman propagator built for Klein-Gordon with static tails","feed_subtitle":"Fredholm inverse on blown-up spacetime yields the microlocal Hadamard condition for massive quantum fields.","key_machinery":"The load-bearing machinery is the 3sc-calculus, a pseudodifferential calculus on the radially compactified spacetime blown up at the two poles where an asymptotically static potential fails to be smooth. Its principal symbol has four components: the standard fiber symbol, the spacetime-infinity symbol, and two indicial families $Œ hat{N}_{\\mathrm{ff},\\pm}(P_V)(\\tau) = \\tau^2 - H_{V_\\pm}$ that encode the limiting spatial Hamiltonians. The argument also uses microlocal cutoffs $Q_{\\mathrm{src}}$ localizing to the radial sources $R_{\\mathrm{src}}$ of the Hamiltonian flow, together with propagation estimates of three types: elliptic estimates away from the characteristic set, real principal-type estimates on the characteristic set away from the radial set, and above- and below-threshold radial point estimates near the sources and sinks. These are assembled into a global Fredholm estimate that yields the Feynman spaces and the invertibility of the map.","core_discovery":"The central claim is that the Feynman propagator for $P_V$ can be realized as the inverse of a Fredholm map between weighted scattering Sobolev spaces, with the Feynman condition encoded by microlocal cutoffs at the radial sources of the Hamiltonian flow rather than by variable-order weights. Theorem 5.1 states that, when the limiting spatial Hamiltonians $H_{V_\\pm} = \\Delta + m^2 + V_\\pm$ have purely absolutely continuous spectrum near $m^2$ and no bound states, the map $P_V : X^{s,\\ell_0,\\ell_+} \\to Y^{s,\\ell_0,\\ell_+}$ is Fredholm; if $V$ is self-adjoint with $V \\in \\rho_{\\mathrm{mf}} \\mathrm{Diff}^1_{3\\mathrm{sc}}$, the map is invertible. Theorem 7.3 extends invertibility to the case of finitely many bound states in $(0,m^2)$ provided $H_{V_\\pm} > 0$. In both settings the inverse $(P_V)^{-1}_{\\mathrm{Fey}}$ is independent of the choice of Sobolev parameters and microlocal cutoffs, and Theorem 6.3 proves the wavefront-set inclusion $\\mathrm{WF}_{\\mathrm{cl}}((P_V)^{-1}_{\\mathrm{Fey}} f) \\subset \\mathrm{WF}_{\\mathrm{cl}}(f) \\cup \\bigcup_{s \\geq 0} \\Phi^s(\\mathrm{WF}_{\\mathrm{cl}}(f) \\cap \\mathrm{Char}(P_0))$, which is the microlocal Hadamard condition.","pith_inferences":["Beyond the paper: the same Fredholm framework should produce a global Feynman parametrix for first-order perturbations and for metric perturbations with weaker decay, since the proof already accommodates potentials of order $r \\geq \\max\\{1,\\ell_+ - \\ell_0\\}$.","Beyond the paper: the assumption that $0 \\notin \\sigma(H_{V_\\pm})$ hints that a zero eigenvalue of the limiting Hamiltonian would generate threshold solutions with linear time growth, which would require a modified Fredholm problem rather than the one constructed here.","Beyond the paper: the wavefront-set inclusion at the level of distributions is a natural first step toward proving the Hadamard property for the associated two-point function, but proving positivity of the resulting bi-solution would require additional analysis that the paper does not carry out."],"forward_implications":["A global Feynman inverse for the massive Klein-Gordon operator exists for asymptotically static potentials without using variable-order Sobolev spaces.","The constructed inverse satisfies the microlocal Hadamard condition, so it qualifies as a distinguished parametrix in the sense used in quantum field theory on curved spacetimes.","The Fredholm setup requires only non-trapping dynamics in the finite region and control of bound states at timelike infinity, not global spectral assumptions on time slices.","In the self-adjoint setting with positive limiting Hamiltonians, the Feynman propagator is uniquely defined and independent of all construction choices.","An anti-Feynman propagator is obtained by interchanging the roles of sources and sinks in the same Fredholm construction."],"supporting_citations":[{"why":"Supplies the dynamical structure theorem on global sources and sinks, the propagation estimates in the 3sc-calculus, and the causal-propagator invertibility used to eliminate kernels and cokernels.","marker":"[2]"},{"why":"Introduces the 3sc-calculus, its principal symbol and indicial operators, and the Fredholm framework for propagators.","marker":"[24]"},{"why":"Develops the many-body propagation-of-singularities estimates that underpin the radial point analysis.","marker":"[25]"},{"why":"Provides the microlocal elliptic regularity and the blueprint for obtaining a Feynman propagator as an inverse of a Fredholm map.","marker":"[28]"},{"why":"Supplies the general method, in the asymptotically hyperbolic and Kerr-de Sitter setting, of realizing propagators as inverses of Fredholm operators.","marker":"[26]"},{"why":"Gives earlier constructions of the massive Feynman propagator on asymptotically Minkowski spacetimes for scattering potentials, the comparison point for the present generalization.","marker":"[9]"},{"why":"Extends the earlier Feynman-propagator construction and is used as a reference for the microlocal Hadamard condition and uniqueness.","marker":"[10]"},{"why":"Establishes the equivalence between the wavefront-set condition and the Hadamard condition, which motivates the definition of the Feynman propagator used here.","marker":"[22]"},{"why":"Supplies the classical distinguished-parametrix construction and the wavefront-set calculus for Feynman parametrices that Theorem 6.3 builds on.","marker":"[6]"}],"fun_headline_variants":["Fredholm inverse realizes Feynman propagator for massive Klein-Gordon","Klein-Gordon Feynman propagator via microlocal cutoffs on compactified spacetime","Klein-Gordon Feynman propagator: Fredholm invertibility and Hadamard condition","Scattering Sobolev spaces yield Klein-Gordon Feynman propagator","Microlocal Hadamard condition for Klein-Gordon Feynman propagator"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The construction rests on the dynamical assumption that the background metric is non-trapping and that the Hamiltonian flow of the unperturbed operator splits its radial set cleanly into global sources and sinks; if trapped bicharacteristics existed or the radial components interacted, the propagation estimates that assemble the Fredholm problem could fail.","fun_headline_variants_meta":{"raw":{"variants":["Fredholm inverse realizes Feynman propagator for massive Klein-Gordon","Klein-Gordon Feynman propagator via microlocal cutoffs on compactified spacetime","Klein-Gordon Feynman propagator: Fredholm invertibility and Hadamard condition","Scattering Sobolev spaces yield Klein-Gordon Feynman propagator","Microlocal Hadamard condition for Klein-Gordon Feynman propagator"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000703,"raw_usage":{"total_tokens":3206,"prompt_tokens":1016,"completion_tokens":2190,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":632,"completion_tokens_details":{"reasoning_tokens":2081}},"tokens_in":632,"tokens_out":2190,"duration_ms":17244,"temperature":1.0,"reasoning_tokens":2081,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T20:47:51.384004+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Check a concrete asymptotically Minkowski metric whose compactified Hamiltonian flow admits a trapped bicharacteristic; if the global Fredholm estimate of Lemma 5.2 still holds, the framework is more general, whereas if it fails, this paper's Feynman construction does not cover that spacetime. Also compute the indicial family $\\tau^2 - H_{V_\\pm}$: if $0$ is an eigenvalue of either limiting Hamiltonian, the main invertibility theorems do not apply.","supporting_citations":[{"cited_title":"Vasy,Propagation of singularities in three-body scattering, Astérisque 262 (2000), vi+151 (English, with English and French summaries)","cited_arxiv_id":null,"evidence_quote":"Introduces the 3sc-calculus, its principal symbol and indicial operators, and the Fredholm framework for propagators."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Develops the many-body propagation-of-singularities estimates that underpin the radial point analysis."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the microlocal elliptic regularity and the blueprint for obtaining a Feynman propagator as an inverse of a Fredholm map."},{"cited_title":"Gérard and M","cited_arxiv_id":null,"evidence_quote":"Gives earlier constructions of the massive Feynman propagator on asymptotically Minkowski spacetimes for scattering potentials, the comparison point for the present generalization."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Extends the earlier Feynman-propagator construction and is used as a reference for the microlocal Hadamard condition and uniqueness."},{"cited_title":"Radzikowski,Micro-local approach to the Hadamard condition in quantum field theory on curved space-time, Comm","cited_arxiv_id":null,"evidence_quote":"Establishes the equivalence between the wavefront-set condition and the Hadamard condition, which motivates the definition of the Feynman propagator used here."}],"review_version":1}