{"id":"5b20f775-7391-4358-b771-0a297ba371df","arxiv_id":"2505.19977","paper_version":2,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For any distributional source, the van Hove-Miyatake model is renormalizable and both renormalization schemes yield the same dressed Hamiltonian: the free field second quantization dΓ(ϖ).","lead":"This paper shows that the ultraviolet divergences of the van Hove-Miyatake model, a toy quantum field theory of a fixed source coupled to bosons, can be controlled for any distributional source. It proves that an algebraic renormalization and a Hamiltonian dressing construction produce unitarily equivalent renormalized Hamiltonians, both equal to the free field Hamiltonian dΓ(ϖ).","discovery_kind":"unification","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Prop. 3.9 identifies the dressed vacuum with the algebraic ground state for the wrong sign of the source: with g=v/ϖ, Def. 3.3 uses e^{a(g)} although Eq. (1) gives e^{a(-g)}, so the claimed GNS representation is that of ω_{∞,-v/ϖ}, not ω_{∞,v/ϖ}.","rationale":"The algebraic construction in Section 2 is standard and the unitary equivalence ι_g H_g ι_g^* = dΓ(ϖ) in Prop. 3.8 is mathematically sound once F^g is defined as the completion with respect to ⟨·,·⟩_g. The load-bearing step is the identification with the algebraic ground state, Prop. 3.9. Reading that proposition carefully, the sign of the shift is inconsistent with Eq. (1) and with Def. 2.2. Because the relevant coherent states for distributional g are disjoint, a sign error here is not cosmetic: it would mean the Hamiltonian construction, as written, renormalizes the van Hove model with source −v rather than v. This is a correctness risk in the central claim, not merely a missing cutoff-limit narrative. The reader's identified weakest assumption concerns the absence of a cutoff-removal limit; that is a legitimate interpretational gap, but it does not address the stronger internal inconsistency in Prop. 3.9. Since the defect is repairable by a consistent replacement g → −g, the conditional verdict remains appropriate: the paper should not be accepted as written, but the main method is likely salvageable. The concrete test above settles whether the sign defect is real or an artifact of the printed conventions.","tokens_in":9932,"tokens_out":30386,"duration_ms":309950,"concrete_test":"Recompute Prop. 3.9 from Definitions 3.3 and 3.7, keeping track of adjoints: set ⟨f,g⟩ = ∫ overline f g, ι_g ε_g(h) = e^{⟨h,g⟩} ε_0(h), and π_0(W(f))ε_0(h) = e^{-π²‖f‖²/2 + iπ⟨f,h⟩}ε_0(h+iπf). Derive the vacuum expectation ⟨ε_g(0),π_g(W(f))ε_g(0)⟩_g. If it is e^{-π²‖f‖²/2 + 2iπ⟨f,g⟩}, rather than e^{-π²‖f‖²/2 − 2πi Re⟨f,g⟩}, the proposition is false as stated. A concrete instance: take d=1, ϖ=1, and real g∈L² with g≠0; evaluate both expressions at f = λ χ for a real compactly supported χ. The phases differ by e^{+2iπλ⟨χ,g⟩} versus e^{−2iπλ⟨χ,g⟩}. The same computation should be repeated after replacing every g in Def. 3.3 and Prop. 3.9 by −g; if Prop. 3.9 then matches Def. 2.2, the required correction is a single global sign change.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central comparison between the algebraic and Hamiltonian approaches is Proposition 3.9. The sign of the shift is inconsistent inside Section 3. For regular v, Eq. (1) defines the dressed scalar product through e^{a(-v/ϖ)} = e^{-a(g)}; but after the definition g:=v/ϖ, Definition 3.3 sets ⟨ψ,φ⟩_g = ⟨e^{a(g)}ψ, e^{a(g)}φ⟩_F, i.e. e^{+a(g)}. This sign propagates: Prop. 3.7(iii) gives ι_g ε_g(f) = e^{⟨f,g⟩} ε_0(f), and the Weyl phase in Prop. 3.9 contains +iπ⟨f,g⟩. Using ordinary sesquilinear Fock inner products, the vacuum expectation of π_g(W(f)) therefore contains a factor e^{+2iπ⟨f,g⟩}, not the e^{-2πi Re⟨f,g⟩} required by the algebraic ground state ω_{∞,g} of Def. 2.2. Equivalently, the GNS representation obtained is that of the state with source −v, not +v. When v/ϖ∉L², ω_{+,v/ϖ} and ω_{−,v/ϖ} are disjoint, so this is not a harmless convention: as written, Prop. 3.9 fails for every distributional source for which the construction is nontrivial. The reader's cutoff-limit concern is real but secondary; this sign defect strikes the theorem itself as stated.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the ultraviolet renormalization of the van Hove–Miyatake scalar field for arbitrary distributional sources v ∈ D'. It develops two constructions: an algebraic one based on explicit (τ,β)-KMS and ground states of the Weyl-algebra dynamics (Section 2), and a Hamiltonian one based on a non-unitary Glimm–Ginibre–Velo dressing (Section 3). The main claim is that the dressed Hilbert space F^g and dressed Hamiltonian H_g are unitarily equivalent to the free second-quantized operator dΓ(ϖ) (Proposition 3.8) and that the Fock vacuum of F^g represents exactly the algebraic ground state ω_{∞,g} of Definition 2.2 (Proposition 3.9).","tokens_in":10352,"tokens_out":14854,"duration_ms":131745,"significance":"If the central equivalence is correct, the paper provides a complete and explicit control of ultraviolet renormalization for the van Hove–Miyatake model for all distributional sources, with no fitted parameters and with a genuine unitary equivalence between an algebraic and an operator-theoretic renormalization scheme. The explicit formulas for the dressed scalar product, the embedding ι_g, and the Weyl representation are valuable and the proofs of Propositions 3.7 and 3.8 are short but essentially complete. The conclusion that the renormalized model is a free field, i.e., that the vHM model is fundamentally trivial for any source, is a clean falsifiable statement. However, the sign mismatch identified below affects the identification with the algebraic ground state precisely in the distributional regime that motivates the paper, so the central claim as written needs correction.","major_comments":[{"comment":"There is a sign inconsistency in the dressing. In Proposition 3.1, with g := v/ϖ, the right-hand side of Eq. (1) is ⟨e^{-a(g)}ϕ, e^{-a(g)}ψ⟩_F. Definition 3.3 instead defines ⟨ψ,ϕ⟩_g = ⟨e^{a(g)}ψ, e^{a(g)}ϕ⟩_F. This sign propagates to the Weyl phase in Proposition 3.9: the resulting vacuum expectation is e^{-1/2||f||² + 2πi Re⟨f,g⟩}, whereas the algebraic ground state ω_{∞,g} of Definition 2.2 has Fourier transform e^{-π²/2||f||² + 2πi Re⟨f,-g⟩}. For g ∉ L²(R^d), these two regular states are disjoint, so Proposition 3.9 as stated fails for every distributional source for which the construction is nontrivial. The proof can be repaired by setting g := -v/ϖ throughout Section 3, or equivalently by replacing e^{a(g)} with e^{-a(g)} in Definitions 3.3 and 3.5 and in the definition of π_g, and the sign must then be tracked consistently through Eq. (1), Proposition 3.7(iii), and Proposition 3.9.","section":"Section 3, Definition 3.3 and Proposition 3.9"},{"comment":"The paper presents the dressed scalar product and dressed Hamiltonian as the outcome of a renormalization procedure, but no cutoff limit is performed. After deriving Eqs. (1)–(2) for v ∈ L², the text says 'We take the right hand side of Eqs. (1) and (2) as the definition' for g ∈ D′. If the renormalized model is intended to be the limit of the regularized GGV construction with cutoffs v_Λ → v, this identity must be proved; otherwise the unitarily equivalent model in Proposition 3.8 describes the paper's own construction rather than a renormalization limit. Since the abstract and introduction explicitly invoke removal of cutoffs, this gap should be addressed, either by supplying the limit theorem or by restating the claim as a definition.","section":"Section 3, Eqs. (1)–(2) and the surrounding text"}],"minor_comments":[{"comment":"The verification that the noncommutative Fourier transforms of the KMS states are quantum positive definite and continuous on finite-dimensional subspaces is left to the reader; because these properties are needed for ω_β to be a regular state, a proof or a precise reference should be supplied.","section":"Definition 2.2"},{"comment":"The scalar product formula ⟨ε_g(h), ε_g(f)⟩_g = e^{⟨h,g⟩ + ⟨f,g⟩ + ⟨h,f⟩} is missing the complex conjugate on the first exponent; it should read e^{\\overline{⟨h,g⟩} + ⟨f,g⟩ + ⟨h,f⟩} if ⟨·,·⟩_2 is sesquilinear in the first argument as defined in Section 2.","section":"Before Proposition 3.9"},{"comment":"The verification that π_g is a *-homomorphism from the Weyl algebra to the unitaries on F^g is left to the reader; since this is a necessary consistency check for the claimed Weyl representation, a short verification should be included.","section":"Proposition 3.9"},{"comment":"There are several typographical errors such as 'Haussdorff' for 'Hausdorff' in Section 3 and 'UL TRA VIOLET' in the title; these should be corrected.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"The sign inconsistency in Proposition 3.9 appears to be a local, fixable slip rather than a defect in the overall algebraic framework; changing g to -v/ϖ in Section 3 would align the Hamiltonian and algebraic constructions. I therefore recommend major revision rather than rejection. The cutoff-limit issue is a genuine gap in the interpretation of the construction as a renormalization, and it should be addressed explicitly."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"I read this carefully. The genuinely new part is the explicit construction of a dressed Hilbert space and renormalized Hamiltonian for arbitrary distributional sources via the non-unitary GGV dressing, together with the proof that this dressed Hamiltonian is unitarily equivalent to dΓ(ϖ). That part is sound: the scalar product is well defined because e^{a(g)} is injective on F_fin(D), and Proposition 3.8 is clean and correct.\n\nThe soft spot is the claimed identification with the algebraic ground state. Definition 3.3 sets the dressed scalar product with e^{a(g)}, where g := v/ϖ. But the regular-source identity in Equation (1), which the paper explicitly takes as its definitional starting point, gives e^{a(−v/ϖ)} = e^{−a(g)}. So the sign is flipped. The flip propagates: in Proposition 3.9 the vacuum expectation of π_g(W(f)) is computed as e^{−1/2||f||^2} e^{2πi Re⟨f,g⟩}, which is the state for source −v. The algebraic ground state of Definition 2.2 has e^{−2πi Re⟨f,g⟩}. For g ∉ L², these two coherent states are disjoint, so Proposition 3.9 fails for exactly the distributional sources the paper targets. This is a load-bearing flaw in the central equivalence claim, though it looks like a fixable sign error rather than a conceptual one: replacing e^{a(g)} with e^{a(−g)} in Definition 3.3 (and adjusting the subsequent signs) makes the theorem work, and none of the operator-theoretic results (Lemma 3.2, Proposition 3.8) are affected.\n\nTwo softer issues. First, the introduction promises a cutoff-removal limit with self-energy and mass renormalization, but Section 3 never does a cutoff limit; it simply takes the right-hand sides of Equations (1) and (2) as definitions. That is a legitimate direct construction, but the paper should say so explicitly instead of implying a regularized limiting procedure. Second, the quantum positive definiteness of the KMS states and the *-homomorphism property of π_g are left to the reader; both are checkable and likely true, but they are part of the proof and should be sketched.\n\nThe algebraic framework is elegant, the citation pattern is honest, and Proposition 2.6 is clearly imported from Arai. The paper would benefit from a referee's attention, principally for the sign issue. I would send it to review, but I would ask the referee to verify the sign in Section 3 and to insist that the direct-definition nature of the renormalization be stated plainly.","headline":"Useful and mostly clean Hamiltonian construction for distributional sources, but the central identification with the algebraic ground state has a sign error: as written, Proposition 3.9 matches the source −v, not +v.","tokens_in":10843,"tokens_out":7068,"would_cite":true,"duration_ms":72437,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["81T16","81R15","81T10"],"pacs":[],"model":"deepseek-v4-flash","headline":"For every distributional source $v\\in D'$, the renormalized van Hove–Miyatake Hamiltonian is unitarily equivalent to the free-field second quantization $d\\Gamma(\\varpi)$, and the algebraic and Hamiltonian constructions coincide.","keywords":["van Hove–Miyatake model","ultraviolet renormalization","distributional source","dressing transformation","non-Fock representation","Weyl algebra","ground states","free field"],"falsifier":"Take a singular source $v$ and a smooth cutoff sequence $v_n\\to v$; compute the limit as $n\\to\\infty$ of the regularized dressed matrix elements $\\langle e^{a^*(-v_n/\\varpi)}\\phi,\\,(H_{\\mathrm{vHM},n}+\\|\\varpi^{-1/2}v_n\\|_2^2)e^{a^*(-v_n/\\varpi)}\\psi\\rangle / \\|e^{a^*(-v_n/\\varpi)}\\Omega_F\\|^2$. If for some $\\phi,\\psi$ this limit differs from $\\langle \\iota_g\\phi,\\,d\\Gamma(\\varpi)\\iota_g\\psi\\rangle$, the direct definition of the dressed Hamiltonian would not reproduce the renormalization limit.","tokens_in":9659,"feed_emoji":"⚛️","tokens_out":17733,"duration_ms":139879,"temperature":0.7,"pith_summary":"The van Hove–Miyatake model couples a fixed, possibly very singular source to a bosonic quantum field; for sources that are not square-integrable the standard Fock-space Hamiltonian is divergent and no unitary dressing can absorb the singularity. This paper establishes that the model is nevertheless renormalizable for every distributional source $v\\in D'$, provided one changes the representation of the canonical commutation relations to a non-Fock one. It offers two independent routes to the same answer: an algebraic construction of ground states of the van Hove dynamical map, and a Hamiltonian construction based on a non-unitary dressing transformation. The main result is that the dressed Hamiltonian $H_g$ on the dressed Hilbert space $\\mathcal{F}^g$ is unitarily equivalent to the free-field Hamiltonian $d\\Gamma(\\varpi)$, and that this representation is exactly the GNS representation of the algebraic ground state. The ultraviolet problem of this model is therefore completely solved, and the renormalized model is a free field for any distributional source.","feed_headline":"Even distributional sources: the renormalized field stays free","feed_subtitle":"By redefining the inner product, ultraviolet divergences disappear and the renormalized model becomes exactly a free field.","key_machinery":"The load-bearing object is the GGV dressing transformation: the non-unitary exponential $e^{a^*(-g)}$, with $g=v/\\varpi$, used in the regular case to conjugate the singular Hamiltonian after subtracting the self-energy. For regular sources the paper records two identities: the dressed inner product is $\\langle e^{a^*(-g)}\\phi,e^{a^*(-g)}\\psi\\rangle/\\|e^{a^*(-g)}\\Omega_F\\|^2=\\langle e^{a(g)}\\phi,e^{a(g)}\\psi\\rangle$, and the dressed Hamiltonian identity is the analogous relation with $d\\Gamma(\\varpi)$ on the right. The paper's move is to take the right-hand sides of these identities as definitions of the dressed scalar product and dressed Hamiltonian for arbitrary $g\\in D'$. Injectivity of $e^{a(g)}$ on the finite-particle subspace gives the new inner product, continuity gives a unitary $\\iota_g$, and closability of the form $q_g(\\phi,\\psi)=\\langle e^{a(g)}\\phi,d\\Gamma(\\varpi)e^{a(g)}\\psi\\rangle_F$ gives a self-adjoint operator $H_g$. The identity $\\iota_g H_g\\iota_g^*=d\\Gamma(\\varpi)$ then reduces the renormalized model to a free field.","core_discovery":"The paper's central claim is that the ultraviolet-singular van Hove–Miyatake model with any distributional source $v\\in D'$ has a well-defined dressed Hamiltonian $H_g$, with $g=v/\\varpi$, and that $H_g$ is unitarily equivalent to the free second quantization $d\\Gamma(\\varpi)$ through the unitary $\\iota_g\\colon \\mathcal{F}^g\\to \\mathcal{F}$. The dressed Hilbert space is obtained by completing the finite-particle subspace in the inner product $\\langle\\psi,\\phi\\rangle_g=\\langle e^{a(g)}\\psi,e^{a(g)}\\phi\\rangle_F$, and the dressed Hamiltonian is defined by the closure of the quadratic form $q_g(\\phi,\\psi)=\\langle e^{a(g)}\\phi,d\\Gamma(\\varpi)e^{a(g)}\\psi\\rangle_F$. The key identity $\\iota_g H_g \\iota_g^*=d\\Gamma(\\varpi)$ makes the renormalized model a free field. The paper further shows, in Propositions 3.8 and 3.9, that the vacuum vector $\\epsilon_g(0)$ in the dressed representation reproduces the expectation values of the algebraic coherent ground state $\\omega_{\\infty,g}$, whose Fourier transform is $e^{-\\pi^2\\|f\\|_2^2/2}\\,e^{2\\pi i\\operatorname{Re}\\langle f,-v/\\varpi\\rangle_2}$, so the two constructions coincide.","pith_inferences":["Not directly asserted in the paper, the same dressing mechanism suggests that a successful ultraviolet renormalization in models with higher-order or spin-coupling interactions may likewise require moving to a dressed, non-Fock representation; the paper signals spin-boson models as the next test case.","A natural testable extension is to repeat the construction for massless bosons or for sources in larger distribution spaces than $D'$, where the exponential-vector argument may need modification.","Since the finite-temperature Gibbs states have explicit formulas, one could check whether dressed Hamiltonians at inverse temperature $\\beta$ converge to $d\\Gamma(\\varpi)$ as $\\beta\\to\\infty$, confirming the zero-temperature limit from the Hamiltonian side.","In this quadratic model the whole content of ultraviolet renormalization is a change of representation: divergent counterterms are absorbed by switching to a dressed Hilbert space rather than by adding operators; whether this remains true for any interaction linear in creation and annihilation operators is an open question."],"forward_implications":["For every distributional source $v\\in D'$, the dressed renormalized Hamiltonian is unitarily equivalent to the free bosonic Hamiltonian $d\\Gamma(\\varpi)$; no interacting part survives renormalization.","Whenever $v/\\varpi\\notin L^2$, the dressing produces a representation of the CCR algebra inequivalent to the Fock representation, so renormalization necessarily moves outside the standard Fock space.","The algebraic ground state $\\omega_{\\infty,g}$ and the Hamiltonian ground state in the dressed space coincide, giving two independent routes to the same renormalized model.","The undressed van Hove Hamiltonian is well defined only when $v/\\varpi\\in L^2$; for more singular sources only the dressed version exists.","The vHM model is fundamentally trivial for any distributional source: its renormalized dynamics is the second quantization of the one-particle dispersion $\\varpi$."],"supporting_citations":[{"why":"Introduces the original van Hove model of a fixed source coupled to a scalar field, whose ultraviolet divergence the paper renormalizes.","marker":"[VH52]"},{"why":"Gives the companion analysis of the same model, providing the divergence problem that the paper's construction addresses.","marker":"[Miy52]"},{"why":"Supplies the theory of inequivalent CCR representations and coherent states used to identify the GNS representation of the algebraic ground state.","marker":"[Ara20]"},{"why":"Provides the standard treatment of van Hove Hamiltonians and the $v/\\varpi\\in L^2$ renormalizability criterion that the paper extends to distributions.","marker":"[Der03]"},{"why":"Establishes that the algebraic van Hove dynamics agrees with the usual Fock dynamics for regular sources, bridging the algebraic and Hamiltonian pictures.","marker":"[FF24]"},{"why":"Gives the Hamiltonian renormalization of quadratic interactions whose dressing identities are the starting point of Section 3.","marker":"[GV70]"},{"why":"Introduces the dressing-transformation technique for boson-field renormalization on which the GGV construction is based.","marker":"[Gli68]"},{"why":"Provides the theorem that weak-* limits of Gibbs states are ground states, used to identify $\\omega_\\infty$ as the algebraic ground state.","marker":"[ST71]"},{"why":"Supplies Nelson's analytic vector theorem, used to show the finite-particle subspace is a core for the free Hamiltonian.","marker":"[RS75]"},{"why":"Establishes the bijection between regular states and quantum positive definite functions that defines the algebraic Gibbs states.","marker":"[Seg59, Seg61]"}],"fun_headline_variants":["Two renormalization routes to a free field, proven equal","Any distributional source renormalizes to a free field","Algebraic vs Hamiltonian renormalization: same free field","Dressed Hamiltonian equals free field for distributional sources"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The construction assumes that the algebraic identities proved for square-integrable sources can simply be taken as definitions for arbitrary distributional sources, so the dressed Hamiltonian is defined directly by the final formula rather than as a limit of regularized cutoffs.","fun_headline_variants_meta":{"raw":{"variants":["Two renormalization routes to a free field, proven equal","Any distributional source renormalizes to a free field","Algebraic vs Hamiltonian renormalization: same free field","Dressed Hamiltonian equals free field for distributional sources"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000965,"raw_usage":{"total_tokens":4086,"prompt_tokens":905,"completion_tokens":3181,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":521,"completion_tokens_details":{"reasoning_tokens":3114}},"tokens_in":521,"tokens_out":3181,"duration_ms":21330,"temperature":1.0,"reasoning_tokens":3114,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T14:03:19.162389+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a singular source $v$ and a smooth cutoff sequence $v_n\\to v$; compute the limit as $n\\to\\infty$ of the regularized dressed matrix elements $\\langle e^{a^*(-v_n/\\varpi)}\\phi,\\,(H_{\\mathrm{vHM},n}+\\|\\varpi^{-1/2}v_n\\|_2^2)e^{a^*(-v_n/\\varpi)}\\psi\\rangle / \\|e^{a^*(-v_n/\\varpi)}\\Omega_F\\|^2$. If for some $\\phi,\\psi$ this limit differs from $\\langle \\iota_g\\phi,\\,d\\Gamma(\\varpi)\\iota_g\\psi\\rangle$, the direct definition of the dressed Hamiltonian would not reproduce the renormalization limit.","supporting_citations":[],"review_version":1}