{"id":"693e3206-bfca-4b8e-a8d6-97f1127ae5fd","arxiv_id":"2412.09708","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For negative coupling, the renormalized non-relativistic and semi-relativistic Nelson semigroups are positivity improving with respect to the Fröhlich cone at every total momentum, proven via Feynman-Kac functional integration.","lead":"This paper proves that the semigroups generated by the ultraviolet-renormalized Nelson Hamiltonians are positivity improving, hence ergodic, for every total momentum, in both the non-relativistic and semi-relativistic models. The proof is a functional-integral argument, and the semi-relativistic case is new.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 1 rests on Feynman–Kac and moment bounds from [MM18]/[HM23] that must hold for all P and for m=0; non-verification of these hypotheses is the main unresolved risk.","rationale":"The central claim is the positivity improving property of the renormalized semigroup. The proof is a reduction to Theorem 8, whose foundation is the Feynman–Kac formula. If that formula or the exponential moment bounds fail for some allowed total momentum or for massless bosons, the positivity argument has no basis. This is the least secure condition because it is imported from earlier works and the paper does not reproduce the hypotheses. The uniformity issue in (3.36)–(3.37) is secondary and repairable for the specific Ω_n used in the application, so it does not by itself overturn the claim. The reader's verdict of CONDITIONAL is appropriate: acceptance should require verification that the cited theorems indeed cover the full parameter range and that the limit interchange is justified.","tokens_in":26767,"tokens_out":26042,"duration_ms":221416,"concrete_test":"Check the statements of [MM18, Thm 7.6] and [HM23, Thm 7.4] and the moment bounds [MM18, Cor 3.21, Thm 4.9] and [HM24a, Thms 6.6–6.7, Lem B.1] to confirm they hold for all P ∈ R^d and for m=0 in the non-relativistic case. Also, in the proof of Theorem 8, replace the appeal to (3.35) by an explicit argument that for each bounded Θ there exists N such that Θ⊂Ω_n for all n≥N; if this is not true for the chosen Ω_n, the limit interchange (3.36)–(3.37) fails.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Theorem 1 is obtained by checking Theorem 8's hypotheses, and Theorem 8's proof is built on the functional-integral representation (2.25) and the bounds (2.26)–(2.28), (3.31a,b) imported from [MM18, Thm 7.6], [HM23, Thm 7.4], [HM24a, Thms 6.6–6.7, Lem B.1]. These results must cover every total momentum P ∈ R^d and, in the non-relativistic case, the massless boson case m=0; otherwise e^{-tH_#(P)} is not identified with the expression whose positivity is proven. The paper neither re-derives these bounds nor states their precise hypotheses, so the applicability over the full parameter range is an assertion rather than a demonstrated fact. A second, internal gap appears in the passage from (3.36) to (3.37): the claimed uniformity in N of the convergence as n→∞ is not a consequence of (3.35), which only bounds the norms; the argument needs the additional fact that for the balls Ω_n={|k|<n} a bounded Θ is eventually contained in Ω_n. Without this, the limit interchange is unjustified. Both defects are repairable, but the first is load-bearing: a failure of the external inputs at any allowed parameter would invalidate Theorem 1.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper claims that for the ultraviolet-renormalized translation-invariant Nelson Hamiltonian, both in the three-dimensional non-relativistic and the two-dimensional semi-relativistic case, the semigroup e^{-tH_{\\#}(P)} is positivity improving with respect to the Fröhlich cone for every total momentum P, provided the coupling constant is negative (Theorem 1). The proof is functional-analytic and probabilistic: it establishes a general Feynman-Kac-type functional-integral calculus (Sections 3.1-3.3), proves positivity of such integrals in progressively more singular situations (Theorems 2, 3, and 8), and then applies Theorem 8 to the renormalized Nelson semigroup using previously established Feynman-Kac representations and moment bounds from [MM18], [HM23], and [HM24a]. The authors also derive a new Trotter product formula (Proposition 6) via a Chernoff-type argument and use it to handle the self-energy renormalization. The paper explicitly notes that the semi-relativistic case at arbitrary total momentum is new, while the non-relativistic case has recent proofs by Miyao and Lampart.","tokens_in":26865,"tokens_out":21730,"duration_ms":208541,"significance":"If the proof is correct, Theorem 1 settles a natural open case: ergodicity, and hence uniqueness and strict positivity of maximal eigenfunctions by the Perron-Frobenius-Faris theorem, for the renormalized semi-relativistic Nelson model at arbitrary total momentum; it also gives a transparent, unified proof for the non-relativistic model. The paper's strategy has real virtues: the core positivity in Theorem 2 is derived from an explicit nonnegative integral kernel and the elementary positivity of E[e^{-iR}] via (3.2), with no fitted parameters and no circular use of the desired conclusion. The limitation statements are honestly placed: Remark 3.5 acknowledges the bounded-domain restriction in the Trotter step, and the dependence on external Feynman-Kac inputs is visible in Section 2.5. The main concerns are that the proof of Theorem 8 contains a nontrivial limit interchange that is not justified as written, and that the paper does not spell out the exact hypotheses under which the imported Feynman-Kac and moment bounds cover the required parameter range.","major_comments":[{"comment":"The passage from (3.36) to (3.37) is not justified as written. From (3.34) and (3.35) one obtains, for each fixed N, norm convergence of the Trotter product as n -> infinity, but the natural telescoping estimate gives an error proportional to N times the per-factor error, so (3.35) does not imply uniformity in N. The subsequent interchange of the limits n -> infinity and N -> infinity is therefore a genuine gap. The gap is repairable within the paper's own framework: since Theta is bounded, for all sufficiently large n one has Theta subset Omega_n, and then Q^Omega_Theta S(n,0) Q^Omega_Theta equals Q^Omega_Theta S(infinity,0) Q^Omega_Theta on F(Theta), with the analogous statement for S(n,k); using this eventual containment, (3.37) follows directly from the n-large case of (3.36), without a double-limit argument. This repair should be written out explicitly, because (3.37) is the identity on which the positivity-improving conclusion of Theorem 8 rests.","section":"Section 3.3, Eqs. (3.36)-(3.37)"},{"comment":"Theorem 1 depends on the Feynman-Kac representation (2.25), on the L^p convergences (2.26)-(2.28), and on the exponential moment bounds (3.31a,b), all of which are imported from [MM18, Theorem 7.6], [HM23, Theorem 7.4], and [HM24a, Theorems 6.6-6.7, Lemma B.1]. These external inputs must hold for every total momentum P in R^d and, in the non-relativistic case, also for massless bosons m=0. The manuscript neither states the precise hypotheses of those theorems nor verifies that they cover the full parameter range used in Theorem 1. If any of those bounds or interchanges fails at an allowed parameter, the identification of e^{-tH_#(P)} with the functional integral whose positivity is proven collapses. Please add a precise statement of the hypotheses and either a short verification or an explicit pointer to the theorem in the cited works that covers exactly the required parameter range, including the massless non-relativistic case.","section":"Section 2.5 and Proof of Theorem 1"}],"minor_comments":[{"comment":"The displayed formula for D_t(tau) contains unresolved TeX control sequences ('bracehtipupleft /bracehtipdownright/bracehtipdownleft /bracehtipupright') and should be cleaned up.","section":"Section 3.3, Eq. (3.28d)"},{"comment":"The dispersion is printed as Psi(p) = sqrt(|p|^2+M^2)-M^2; it should be sqrt(|p|^2+M^2)-M, which is the function with Psi(0)=0 and subadditivity used in the text.","section":"Example 3.4(1)"},{"comment":"The symbol f(t) in (3.33) is never defined in the statement of Theorem 8; presumably it is the pointwise limit of f_n(t), and that hypothesis should be stated explicitly.","section":"Theorem 8, Eq. (3.33)"},{"comment":"The assertion that h(t,x) is selfadjoint for arbitrary g_+ and g_- requires an additional condition such as g_-(t)=g_+(t) (or a specified relation between the two test functions); in the Nelson application the two functions are indeed equal, so this is a statement-level gap rather than a flaw in the main application.","section":"Proposition 5, Eq. (3.20)"},{"comment":"The notation P_{Omega_n} psi is used but not defined; Section 2.1 already uses P_n for Fock-space projections, and the projection onto F(Theta) is denoted Q^Omega_Theta in (2.3), so this should be aligned to avoid confusion.","section":"Theorem 3, proof"},{"comment":"The statement 'P in R^d' should specify the dimension for each model: d=3 for #=nr and d=2 for #=sr, since those are the dimensions in which the models are defined.","section":"Theorem 1 and Section 2.2"}],"recommendation":"major_revision","confidential_remarks":"The main theorem is plausible and the proof strategy is coherent, but the limit interchange between (3.36) and (3.37) is a real gap in the written proof of Theorem 8, and the reliance on external Feynman-Kac and moment bounds should be made explicit for the full parameter range. Both issues appear repairable without changing the paper's scope, so I recommend major revision rather than rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nThe headline: this paper proves, via Feynman-Kac and kernel positivity, that the ultraviolet-renormalized non- and semi-relativistic Nelson semigroups are positivity improving with respect to the Fröhlich cone at every total momentum, for negative coupling. The semi-relativistic case is genuinely new; the non-relativistic case was already known (Miyao, Lampart), and the paper says so plainly. The functional-integral route is clean and repackages known positivity arguments in a way that may be reusable for other translation-invariant models.\n\nWhat I checked: Theorem 2's kernel-positivity step is legitimate—(3.2) makes E[e^{-iR}] > 0 when the Lévy symbol is real. The limiting argument in Theorem 3 follows the standard pattern, and the application of Theorem 8 to the Nelson case satisfies (3.32) because P_n = P kills the phase. The paper is careful about attributing prior results.\n\nSoft spots, in order of severity:\n\n1. The passage from (3.36) to (3.37) interchanges n→∞ and the Trotter limit N→∞. The paper says the convergence is uniform in N by (3.35), but (3.35) only bounds norms; it does not by itself give uniformity of the convergence rate in N. This needs an argument—for example, showing directly that S(∞,0) satisfies the flow equation and applying Chernoff to the limiting family. As written, it is a gap, though likely repairable.\n\n2. Proposition 5 asserts selfadjointness of h(t,x) for arbitrary g±. That only holds when the annihilation and creation test functions coincide (in the application they do: g_+ = g_- = -v_n). The statement should be restricted; the main theorem is unaffected.\n\n3. The proof leans heavily on imported Feynman-Kac formulas and exponential moment bounds from [MM18, HM23, HM24a]. That is normal, but the paper does not state the precise hypotheses under which those hold for all P and for massless bosons in the non-relativistic case. A referee will want those spelled out or at least cited with theorem numbers and parameter ranges.\n\nThe internal math that is displayed checks out; the novelty is real for the semi-relativistic case; the limitations are repairable. This deserves a serious referee; I would send it out, asking for the (3.37) justification and the Proposition 5 restriction.\n\nBest,","headline":"A clean functional-integral proof that the renormalized semi-relativistic Nelson semigroup is positivity improving at every momentum; the main theorem is new and mostly sound, but a few technical gaps need patching.","tokens_in":27622,"tokens_out":3607,"would_cite":true,"duration_ms":34662,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["81S40","47D08"],"pacs":[],"model":"deepseek-v4-flash","headline":"Negative coupling makes the semigroups of the ultraviolet-renormalized non-relativistic and semi-relativistic Nelson Hamiltonians positivity improving for every total momentum; the semi-relativistic case at $P \\neq 0$ is new.","keywords":["Nelson model","semi-relativistic Nelson model","positivity improving","Fröhlich cone","ergodic semigroup","Feynman–Kac formula","ultraviolet renormalization","total momentum"],"falsifier":"Exhibit vectors $\\psi,\\varphi$ in the Fröhlich cone and some $P \\neq 0$, $t>0$ with $\\langle \\psi, e^{-tH_{\\mathrm{sr}}(P)}\\varphi\\rangle = 0$; that would directly contradict Theorem 1. A less direct but checkable failure point is the exponential moment bound (3.31b) for the semi-relativistic model: if it diverges for some $P$, the monotone-convergence step in Theorem 8 breaks and the functional-integral proof no longer goes through.","tokens_in":26368,"feed_emoji":"⚛️","tokens_out":12148,"duration_ms":103949,"temperature":0.7,"pith_summary":"The paper establishes Theorem 1: for negative coupling $\\lambda < 0$, the semigroups $e^{-tH_{\\#}(P)}$ generated by the ultraviolet-renormalized non-relativistic ($\\# = \\mathrm{nr}$) and semi-relativistic ($\\# = \\mathrm{sr}$) Nelson Hamiltonians are positivity improving with respect to the Fröhlich cone, for every total momentum $P \\in \\mathbb{R}^d$ and every $t > 0$. Positivity improving means the semigroup maps every nonzero vector in the cone to a vector that has strictly positive inner product with every cone vector. By the Perron–Frobenius–Faris theorem this implies that any maximal eigenvalue of $H_{\\#}(P)$ is non-degenerate and has a unique strictly positive eigenvector. The non-relativistic renormalized case was known through different methods, while the semi-relativistic case at arbitrary total momentum is new; the proof here is a single functional-integral argument covering both cases.","feed_headline":"Negative coupling yields positivity improving Nelson semigroups","feed_subtitle":"Semi-relativistic case now proven at any total momentum, not just P = 0.","key_machinery":"The key object is the Feynman–Kac representation (2.25), which writes the semigroup as an expectation over a Lévy process whose characteristic exponent is the particle dispersion. Positivity is read off from the integral kernels: the assumptions in Theorem 2 make the kernel $F^{\\pi,i,\\ell}_{k,p,t}$ nonnegative almost everywhere, and the Lévy property makes $E[e^{-iR}] > 0$ strictly. For the renormalized models, the new Trotter product formula (Proposition 6), proved via the Chernoff product formula, is what allows the limit from ultraviolet-regularized to renormalized semigroups to preserve positivity improving. The Fröhlich cone is the self-dual cone of Fock-space vectors whose $n$-particle components are nonnegative almost everywhere for every $n$.","core_discovery":"The central claim is that the renormalized Nelson semigroups are ergodic at negative coupling for all total momenta. The proof represents $e^{-tH_{\\#}(P)}$ by the Feynman–Kac formula $e^{-tH_{\\#}(P)} = \\mathbb{E}\\left[ e^{u_t} F^{\\omega}_{t/2}(U^-) F^{\\omega}_{t/2}(U^+)^{*} e^{i(P-\\mathrm{d}\\Gamma(\\hat{p}))\\cdot X_t} \\right]$, and shows that every matrix element against Fröhlich-cone vectors can be written as an integral of a nonnegative kernel times $E[e^{-iR}]>0$, the characteristic function expectation of a Lévy process. The renormalized limit is controlled by convergence of the random objects $u_t, U^\\pm$ in $L^p$ together with exponential moment bounds. A new Trotter product formula (Proposition 6), derived from a Chernoff-type theorem, lets the authors pass from the ultraviolet-regularized to the renormalized semigroup by decomposing Fock space onto bounded momentum subsets, where the quadratic dispersion satisfies the needed lower-boundedness assumption. The outcome is that every maximal eigenvalue of $H_{\\#}(P)$ is simple and the eigenspace is spanned by a unique strictly positive vector.","pith_inferences":["The same positivity machinery should apply to other translation-invariant polaron-type models that admit a Feynman–Kac representation with real characteristic exponent and exponential moment bounds, such as Hamiltonians with subcritical ultraviolet singularities.","The bounded-subset reduction in Proposition 6 is needed only because Assumption A fails for the quadratic dispersion on unbounded sets; a Trotter formula tolerating an unbounded-from-below $L$ would allow $\\Theta = \\Omega$ and a more direct proof of Theorem 8.","A concrete test of the method's limits is the three-dimensional semi-relativistic model, where the paper notes a similar result is not expected; if the required moment bounds fail there, the functional-integral route would mark the boundary of its applicability."],"forward_implications":["For $\\lambda < 0$, both $e^{-tH_{\\mathrm{nr}}(P)}$ and $e^{-tH_{\\mathrm{sr}}(P)}$ are positivity improving and hence ergodic for every total momentum $P$.","Every maximal eigenvalue of $H_{\\mathrm{nr}}(P)$ and $H_{\\mathrm{sr}}(P)$ at negative coupling is non-degenerate, with a unique strictly positive eigenvector.","The proof also covers the ultraviolet-cutoff-free Fröhlich polaron, giving a direct functional-integral proof of its ergodicity for any $P$.","The unified argument removes the earlier restriction to $P = 0$ in the semi-relativistic renormalized case."],"supporting_citations":[{"why":"It supplies the Feynman–Kac representation for the non-relativistic renormalized semigroup and the exponential moment bounds used to verify Theorem 8.","marker":"[MM18]"},{"why":"It supplies the Feynman–Kac representation for the semi-relativistic renormalized semigroup in two dimensions.","marker":"[HM23]"},{"why":"It provides the exponential moment bounds and $U^{\\pm}$ convergence for the semi-relativistic case, cited as Theorems 6.6–6.7 and Lemma B.1.","marker":"[HM24a]"},{"why":"It gives the Chernoff-type product formula for time-dependent operators used to prove the Trotter product formula (Proposition 6).","marker":"[Vui10]"}],"fun_headline_variants":["Ergodic Nelson semigroups proven for all momenta","Semi-relativistic Nelson semigroups now ergodic at any P","Renormalized Nelson semigroups: ergodic at every momentum","All momenta now included in Nelson ergodicity proof","Positivity improving: Nelson semigroups for any P"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof stands on the Feynman–Kac representation (2.25) and on the imported convergence and exponential moment bounds holding uniformly in the total momentum $P$, together with the Trotter product formula of Proposition 6; if any of those imported bounds fails, the functional-integral argument for Theorem 1 collapses.","fun_headline_variants_meta":{"raw":{"variants":["Ergodic Nelson semigroups proven for all momenta","Semi-relativistic Nelson semigroups now ergodic at any P","Renormalized Nelson semigroups: ergodic at every momentum","All momenta now included in Nelson ergodicity proof","Positivity improving: Nelson semigroups for any P"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000393,"raw_usage":{"total_tokens":2030,"prompt_tokens":878,"completion_tokens":1152,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":494,"completion_tokens_details":{"reasoning_tokens":1064}},"tokens_in":494,"tokens_out":1152,"duration_ms":9243,"temperature":1.0,"reasoning_tokens":1064,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T16:53:46.595239+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Exhibit vectors $\\psi,\\varphi$ in the Fröhlich cone and some $P \\neq 0$, $t>0$ with $\\langle \\psi, e^{-tH_{\\mathrm{sr}}(P)}\\varphi\\rangle = 0$; that would directly contradict Theorem 1. A less direct but checkable failure point is the exponential moment bound (3.31b) for the semi-relativistic model: if it diverges for some $P$, the monotone-convergence step in Theorem 8 breaks and the functional-integral proof no longer goes through.","supporting_citations":[],"review_version":1}