{"id":"4a17d281-1d02-42f6-a751-d4f70e82094f","arxiv_id":"1909.01081","paper_version":1,"verdict":"REJECT","confidence":"MODERATE","novelty_score":2.0,"correctness_risk":"high","formal_verification":"none","parameter_count":4,"one_line_summary":"A finite-volume fractional-Laplacian regularization is claimed to make 2D Euclidean scalar field path integrals finite after a specially chosen renormalization, but the central limit argument rests on an unproved determinant expansion.","lead":"Using fractional powers of the Laplacian on a finite box, the paper proposes a functional-integral regularization of Euclidean quantum field theories in which fields are ordinary square-integrable functions. It argues that with a specially chosen bare coupling the perturbative expansion stays finite as the regulator is removed, but the key steps rely on unproved expansions and previous self-cited work.","discovery_kind":"incremental","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The finite-limit claim rests on the unproved determinant expansion Eq. (22); higher-order terms are uncontrolled and the determinant uses a zero-diagonal covariance inconsistent with the finite-volume covariance of Eq. (15).","rationale":"The central claim is explicitly about a limit of regularized path integrals, and the only step that converts the singular bare-coupling prefactors into finite expressions is the determinant expansion Eq. (22). The reader's weakest_assumption identifies exactly this expansion, and my stress-test sharpens it: the expansion is not merely unproved; as written it is based on a translation-invariant kernel with zero diagonal, while the actual finite-volume covariance of Eq. (15) has a positive, divergent diagonal, and the remainder terms in Eq. (22) are not controlled uniformly in N and v. These are concrete internal correctness risks, not disagreements with external consensus. The standard Wiener-measure material in Section 4 is not connected to the cutoff-removal argument and does not provide independent support for the finite-limit claim. A direct numerical computation of L_N for N=2,3 with the spectral covariance is feasible and would settle whether the advertised cancellation is a theorem or an artifact of an invalid determinant approximation. Because the central gap is the same one identified by the reader and no verification is supplied, the REJECT verdict stands without adjustment.","tokens_in":14313,"tokens_out":15118,"duration_ms":239360,"concrete_test":"Numerically test Eq. (22) for N=2 and N=3 using the finite-volume covariance of Eq. (15). On Omega = [0,L]^2, take Dirichlet eigenvalues lambda_{mn} = pi^2 (m^2+n^2)/L^2 and set G_alpha(x,y) = (2/L)^2 sum_{m,n>=1} (lambda_{mn}^alpha + m0^2)^{-1} sin(m pi x/L) sin(n pi y/L) sin(m pi x'/L) sin(n pi y'/L), with m0^2 small, e.g. m0^2 = alpha - 1. For alpha = 1+epsilon with epsilon = 10^{-2}, 10^{-3}, 10^{-4}, evaluate L_N(alpha,L) = integral_{Omega^N} det^{-1/2}(G_alpha(x_i,x_j)) d^N x by deterministic quadrature. Compare L_N(alpha,L) with (1-alpha)^{N/2} L^{2N} (4 pi)^{N/2} (N-1)^{-1/2}. If the ratio does not approach 1, or if the remainder after the leading term, multiplied by g_bare^N and integrated, does not vanish uniformly, then the cancellation in Eq. (24) fails and the central claim is not established.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 3's conclusion, that Eq. (14) has a finite limit at alpha = 1 under Eq. (17) and Eq. (18), is obtained by pairing the divergent prefactor g_bare ~ (1-alpha)^{-1/2} v^{-1} with the determinant expansion Eq. (22), L_N(alpha,v) = (1-alpha)^{N/2} C_N + ... . That expansion is the only step that converts the singular prefactors into finite quantities, and it is unsupported in three concrete ways. First, Eq. (22) is not proved in the paper; it is cited to the author's earlier work. Second, the leading coefficient C_N in Eq. (23) is computed from the matrix A_ii=0, A_ij=1, which comes from a translation-invariant kernel whose diagonal vanishes for alpha > 1. But the Gaussian covariance in Eq. (15), whose trace is computed in Eq. (9), has a positive diagonal; a finite-volume spectral covariance has G_alpha(x,x) = sum_n lambda_n^{-alpha} phi_n(x)^2 ~ c/(alpha-1) as alpha -> 1. A zero-diagonal determinant and a positive-diagonal determinant with the same off-diagonal behavior need not have the same leading coefficient, and the v-scaling of the remainder can change. Third, the higher-order terms in Eq. (22) are discarded without a uniform estimate in N and v. Since the limit in Eq. (18) sends v -> infinity after alpha -> 1, any remainder of the form (1-alpha)^m v^p will survive the limit unless p=0 or m grows with p; no such control is provided. These gaps are load-bearing because without Eq. (22) the advertised cancellation is absent and the finite-limit conclusion is unsupported.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a finite-volume analytic regularization of Euclidean scalar field theory based on fractional powers of the Laplacian. For α > D/2 the Gaussian measure with covariance χ_Ω(−Δ)^{−α}χ_Ω is constructed via the Minlos–Bochner theorem (Section 2). In Section 3 the author defines a bare coupling g_bare(α,v) = g_rem/((1−α)^{1/2} v) in Eq. (17) and claims that the perturbative generating functional Eq. (14), understood through the nested limit in Eq. (18), has a finite limit at α = 1 and v → ∞. The remaining sections review the Wiener measure construction and propose a Feynman–Wiener geometrodynamical propagator, with appendices on distributional Fourier transforms and functional calculus.","tokens_in":14696,"tokens_out":3751,"duration_ms":227861,"significance":"If the central claim were established, the paper would provide a concrete finite-volume regularization in which a 2D scalar interaction has a finite perturbative generating functional, and it would clarify the nature of ultraviolet infinities in a mathematically explicit setting. The first two theorems of Section 2 are standard and correctly identify the range α > D/2 in which the fractional Laplacian inverse is trace class; the bound in Eq. (19) is a real estimate given the determinant expansion. However, the central removal step is not proved: the determinant expansion Eq. (22) is asserted and cited to the author's own previous work, the covariance used to compute its leading coefficient appears inconsistent with the covariance of Eq. (15), and the higher-order terms are discarded without uniform control. No empirical prediction or machine-checked proof is supplied. The result therefore remains an unsupported claim rather than an established theorem.","major_comments":[{"comment":"The entire cancellation that converts the divergent prefactors (1−α)^{−N/2} into the finite expression in Eq. (24) relies on the expansion L_N(α,v) = (1−α)^{N/2} C_N + ⋯, but this expansion is not proved in the manuscript; it is cited to the author's earlier papers [2,3]. Since this expansion is the only mechanism that neutralizes the singular bare coupling of Eq. (17), the central conclusion 'has a finite limit for α = 1' is unsupported.","section":"§3, Eq. (22)"},{"comment":"The leading coefficient C_N in Eq. (23) is computed from the matrix A_ii = 0, A_ij = 1, which corresponds to a translation-invariant kernel whose diagonal vanishes for α > 1. But the actual covariance of the measure in Eq. (15) has a positive diagonal: its trace, computed in Eq. (9), diverges as α → 1, and a finite-volume spectral covariance has G_α(x,x) ∼ c/(α−1). A zero-diagonal determinant and a positive-diagonal determinant with similar off-diagonal behavior need not have the same leading coefficient, and the v-scaling of the remainder can differ. Therefore Eq. (22) is not justified as stated.","section":"§3, Eqs. (15) and (23)"},{"comment":"The higher-order terms in Eq. (22) are discarded without a uniform estimate in N and v. The limit order in Eq. (18) sends v → ∞ after α → 1, so a remainder of the form (1−α)^m v^p would survive the limit unless p = 0 or m grows with p. No such control is provided. Thus even if the leading term of Eq. (22) were correct, the claimed finite limit is not established.","section":"§3, Eqs. (18) and (22)"},{"comment":"The nested limit in Eq. (18) is presented as the 'R.P. Feynman sense', but the proof of Theorem 1 bounds lim_{N→∞} lim_{α→1} |I_N| at fixed v. The order in Eq. (18) instead takes α → 1 before N → ∞ and before v → ∞. No dominated-convergence or uniformity argument is supplied to justify this interchange or even to show that the stated order is well defined. This ambiguity affects the meaning of the central claim.","section":"§3, Eq. (18)"}],"minor_comments":[{"comment":"The text attributes statements to 'S. Coleman ([4])' and 'G. Hoft ([5])', but the listed references [4] and [5] are Green–Schwarz–Witten and Klaiber; the citations should be corrected.","section":"References [4] and [5]"},{"comment":"The manuscript contains many typographical and grammatical errors ('Districutions', 'quntum', 'theories by theirs turn', 'analitically', 'Feynman-Dhyson'), and the equations contain garbled symbols and inconsistent notation; a careful proofread is needed.","section":"Throughout"},{"comment":"The inversion formula in Eq. (C-9) rests on an explicitly non-proved hypothesis and an 'open problem' in infinite-dimensional analysis; since this appendix is not used in the main argument, it should be labeled as formal or speculative.","section":"Appendix C"},{"comment":"The determinant factor for the matrix A is written ambiguously as '|(−1)(N − 1)(−1)^N|'; the sign and absolute value should be stated in a single unambiguous expression, such as |det A| = N − 1 for N ≥ 2.","section":"§3, Eq. (24)"}],"recommendation":"reject","confidential_remarks":"The central claim of Section 3 depends on the determinant expansion Eq. (22), which is not proved here and is cited only to the author's own prior work. The covariance mismatch between Eq. (23) and Eq. (15) raises substantive doubt about the expansion, and the uncontrolled remainders are not a local fix. In my view this is a load-bearing gap that cannot be repaired within the scope of a revision."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a research note that recycles the author's earlier finite-volume fractional-Laplacian regularization and adds a claim about cutoff removal at alpha=1. The new part is mostly the claim, and it does not hold together.\n\nWhat the paper does well: the construction of the Gaussian measure for alpha > D/2 is standard Minlos/Bochner material, and the bound on the truncated perturbative series in (19) is a genuine estimate, but only after you accept Eq. (22). The Wiener measure section is textbook. The 4D bi-Laplacian remark is a sketch, not a result.\n\nWhere it falls apart: the entire finite-limit conclusion in Section 3 rests on the determinant expansion Eq. (22), which is asserted, not proved, and cited to the author's own previous work. The leading coefficient is computed from a matrix with zero diagonal, while the actual covariance in Eq. (15) has a positive diagonal; the two determinants need not have the same leading behavior. The higher-order terms are discarded with no uniform control in N and v, and the limit order in Eq. (18) is never justified. Since the bare coupling is chosen exactly to cancel the (1-alpha)^{N/2} factors, the argument is circular at the key step. The paper also contains a false statement in the proof of Theorem 2 (\"if phi(x)=+infinity for phi in L2\"), and it later labels the finiteness of correlation functions a conjecture after claiming a proof. That is an internal tension the reader should not have to resolve.\n\nThe citation pattern is mostly self-citation; that is not automatically a problem, but here it becomes load-bearing because the main expansion is not reproduced. A referee would have to chase down the author's earlier papers to see if Eq. (22) is even true, and the stress-test note identifies a concrete reason to doubt it.\n\nWho this is for: possibly someone working on the same regularization scheme. Most readers of a general physics journal will not get much from it, and the advertised general lesson about infinities is not established.\n\nRecommendation: I would not send this to peer review in its current form. If the author provides a proof of Eq. (22) with controlled remainders, the finite-limit claim could be worth a serious look.","headline":"A recycling of the author's own fractional-Laplacian regularization, with the finite-limit claim resting on an unproved determinant expansion and a circular coupling choice.","tokens_in":15250,"tokens_out":4414,"would_cite":false,"duration_ms":517710,"reading_group":"no","serious_thinker":"no","would_accept_peer_review":false},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["81T08","81T16","81S40","28C20"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper claims that one coupling rescaling, $g_{\\mathrm{bare}} = g_{\\mathrm{rem}}/((1-\\alpha)^{1/2}v)$, makes a two-dimensional scalar field theory's generating functional finite at the ultraviolet limit $\\alpha = 1$.","keywords":["Euclidean quantum field theory","functional integral","fractional Laplacian regularization","coupling constant renormalization","ultraviolet divergences","probability measures on function spaces","Minlos theorem","finite-volume regularization"],"falsifier":"Take the two-point case on a torus, where the fractional Laplacian has an explicit eigenbasis, and compute $L_2(\\alpha,v) = \\int_{\\Omega^2} d^2x_1\\, d^2x_2\\, \\det^{-1/2}[L^{-1}_\\alpha(x_i,x_j)]$; the cancellation (24) at second order requires $L_2(\\alpha,v) \\to (1-\\alpha)\\,4\\pi v^2$ as $\\alpha \\to 1$, up to terms that vanish after multiplication by $(g_{\\mathrm{bare}})^2$. Repeating the check at third order — where the predicted limit involves $(N-1)^{-1/2}$ with $N=3$ — would test whether the remainder terms genuinely drop out; any mismatch in the power of $(1-\\alpha)$ or a surviving remainder falsifies the finiteness claim.","tokens_in":14062,"feed_emoji":"⚛️","tokens_out":21179,"duration_ms":175163,"temperature":0.7,"pith_summary":"The paper proposes a measure-theoretic cure for ultraviolet infinities in two-dimensional Euclidean scalar quantum field theory and claims to carry out the cutoff removal in one model. The device is a fractional-power Laplacian $(-\\Delta)^\\alpha$ restricted to a finite volume $\\Omega$: for $\\alpha > 1$ the free-field measure is supported on ordinary square-integrable functions, so interactions are well-defined objects instead of formal divergences. The central assertion is that, with the bare coupling rescaled as $g_{\\mathrm{bare}}(\\alpha,v) = g_{\\mathrm{rem}}/((1-\\alpha)^{1/2} v)$, the perturbatively defined generating functional has a finite limit as the regulator $\\alpha \\to 1$ and the volume $v \\to \\infty$ when the limits are taken in the order fixed by eq(18). A sympathetic reader would care because the argument exhibits the renormalization mechanism concretely: the $(1-\\alpha)^{-1/2}$ blow-up in the coupling and the $(1-\\alpha)^{1/2}$ leading behavior of a determinant integral cancel exactly, order by order, inside a framework where the fields are genuine functions.","feed_headline":"One coupling rescaling makes a 2D field theory finite","feed_subtitle":"Rescaling the bare charge by a square-root factor cancels ultraviolet blowups as the regulator approaches one.","key_machinery":"The load-bearing object is the analytically regularized covariance $L^{-1}_{\\alpha,\\Omega,m} = \\chi_\\Omega [(-\\Delta)^\\alpha + m^2]^{-1}\\chi_\\Omega$, which for $\\alpha > D/2$ is trace-class and, by the Minlos–Bochner theorem, defines a probability measure supported on $L^2(\\Omega)$ — ordinary square-integrable functions instead of distributions — so interactions become legitimate objects. The renormalization prescription $g_{\\mathrm{bare}}(\\alpha,v) = g_{\\mathrm{rem}}/((1-\\alpha)^{1/2} v)$ is engineered so that its $(1-\\alpha)^{-1/2}$ blow-up matches, at order $N$, the $(1-\\alpha)^{N/2}$ leading behavior supplied by the determinant expansion (22), $L_N(\\alpha,v) = \\int_{\\Omega^N} d^2x_1\\cdots d^2x_N\\, \\det^{-1/2}[L^{-1}_\\alpha(x_i,x_j)] = (1-\\alpha)^{N/2}C_N + \\cdots$, with $C_N = v^N (4\\pi)^{N/2}(\\det A)^{-1/2}$ and $A$ the matrix that is zero on the diagonal and one off it. The determinant of $A$ contributes the factor $(N-1)^{-1/2}$ that suppresses higher orders, making the perturbative series summable; the author also claims the analogous kernel $(-\\Delta^2)^{-\\alpha}$ carries the same argument in four dimensions.","core_discovery":"On the paper's own terms, the central discovery is the finiteness statement of Section 3: the functional path integral (14), interpreted through the Feynman perturbative definition (18) — limits nested so that $m^2 \\to 0$ first, then $\\alpha \\to 1$, then $N \\to \\infty$, and finally $v \\to \\infty$ — has a finite limit at $\\alpha = 1$ once the bare coupling is set to $g_{\\mathrm{bare}}(\\alpha,v) = g_{\\mathrm{rem}}/((1-\\alpha)^{1/2}v)$. The proof is an estimate: each $N$-th order term is bounded by $|g_{\\mathrm{rem}}|^N (4\\pi)^{N/2}/(N-1)^{1/2}$ times the norm of the interaction's Fourier transform, so the whole series is dominated by $C^2 e^C$ with $C = (4\\pi)^{1/2}|g_{\\mathrm{rem}}|\\|\\tilde{V}\\|_{L^\\infty}$; consequently the generating functional is finite and continuous in the source $j \\in L^2(\\Omega)$. The same construction is claimed to work in four dimensions with the square-Laplacian kernel $(-\\Delta^2)^{-\\alpha}$, while the competing propagator prescription (29), which converges to the ordinary massless two-dimensional Green function, is shown to fail because that Green function is not a tempered distribution.","pith_inferences":["If expansion (22) can be proved with explicit control of its remainder — e.g., by diagonalizing the fractional Laplacian on a torus — the same 'matching powers' rule, pairing a $(1-\\alpha)^{-1/2}$ coupling divergence against a $(1-\\alpha)^{N/2}$ determinant factor, should generalize to other super-renormalizable models with power-like ultraviolet singularities.","The designated order of limits in (18) is load-bearing: the paper sends $m^2 \\to 0$, then $\\alpha \\to 1$, then $N \\to \\infty$, then $v \\to \\infty$, and a numerical check of the low-order coefficients under a reversed order — say $N \\to \\infty$ before $\\alpha \\to 1$ — would show whether the divergence the construction is designed to cancel reappears.","The claimed four-dimensional extension with kernel $(-\\Delta^2)^{-\\alpha}$ is the natural place to look for a counterexample, since the kernel's power in $|x-y|$ differs; computing $L_2$ and $L_3$ there would show whether the determinant cancellation survives in higher dimension.","The contrast between the successful determinant route and the failed naive propagator suggests a general moral the paper leaves implicit: in a measure-based formulation, renormalizability is a property of the pairing between the coupling's regulator dependence and the covariance's singular behavior, not of the bare action alone."],"forward_implications":["The full generating functional (14), under the limit order (18), is finite and continuous in the source $j \\in L^2(\\Omega)$, so the model yields a genuine Euclidean field-theory object rather than a formal divergent series.","Every truncation of the perturbative series obeys the uniform bound $|I_N| \\le C^2 e^C$ with $C = (4\\pi)^{1/2}|g_{\\mathrm{rem}}|\\|\\tilde{V}\\|_{L^\\infty}$, so convergence is dominated by an ordinary exponential series.","For $\\alpha > 1$ and $\\delta > 0$, the exponential-cutoff interactions have finite $N$-point functions (eq 31), so the framework supports non-Gaussian measures before any cutoff is removed.","The same proof is claimed to extend to a four-dimensional model whose kinetic operator is the square Laplacian $(-\\Delta^2)^\\alpha$ on a finite volume in $\\mathbb{R}^4$.","The naive propagator prescription (29) is shown to fail because the massless two-dimensional Green function is not a tempered distribution, so the finite-limit claim is tied to the specific determinant-cancellation route rather than to any propagator that approaches the standard Green function."],"supporting_citations":[{"why":"Supplies the standard framework the paper builds on: analytic continuation of Euclidean fields, the Minlos–Bochner representation of generating functionals, and the P(φ)2 background.","marker":"[1]"},{"why":"The author's earlier paper that provides the fractional-Laplacian regularization, the trace-class measure on the finite box, and the determinant expansion (22) on which the cancellation rests.","marker":"[2]"},{"why":"Provides the determinant evaluation with the off-diagonal-ones matrix that yields the one-over-square-root-of-(N−1) suppression factor in eq(24).","marker":"[3]"},{"why":"Cited for the fact that the massless two-dimensional Green function is not a tempered distribution, which rules out the naive propagator prescription (29).","marker":"[4]"},{"why":"Cited for the Feynman perturbative definition (Chapter 5, §5.2, eq(5.11)) that fixes the order of limits in (18), and for the Wiener–Kac constructions used in Sections 4–5.","marker":"[7]"}],"fun_headline_variants":["Square-root rescaling tames 2D UV infinities","Finite path integral via bare charge rescaling","Coupling rescaling cancels 2D field theory blowups","One rescaling removes 2D QFT infinities"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The finiteness claim rests on two unproved premises: the determinant expansion (22), which is asserted from the author's earlier work and must have leading term $(1-\\alpha)^{N/2}C_N$ with vanishing remainders, and the specific nested order of limits in (18) that identifies the path integral with the $N \\to \\infty$ limit of its perturbative series before the volume diverges.","fun_headline_variants_meta":{"raw":{"variants":["Square-root rescaling tames 2D UV infinities","Finite path integral via bare charge rescaling","Coupling rescaling cancels 2D field theory blowups","One rescaling removes 2D QFT infinities"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000458,"raw_usage":{"total_tokens":2245,"prompt_tokens":844,"completion_tokens":1401,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":460,"completion_tokens_details":{"reasoning_tokens":1333}},"tokens_in":460,"tokens_out":1401,"duration_ms":10842,"temperature":1.0,"reasoning_tokens":1333,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T12:00:29.311707+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take the two-point case on a torus, where the fractional Laplacian has an explicit eigenbasis, and compute $L_2(\\alpha,v) = \\int_{\\Omega^2} d^2x_1\\, d^2x_2\\, \\det^{-1/2}[L^{-1}_\\alpha(x_i,x_j)]$; the cancellation (24) at second order requires $L_2(\\alpha,v) \\to (1-\\alpha)\\,4\\pi v^2$ as $\\alpha \\to 1$, up to terms that vanish after multiplication by $(g_{\\mathrm{bare}})^2$. Repeating the check at third order — where the predicted limit involves $(N-1)^{-1/2}$ with $N=3$ — would test whether the remainder terms genuinely drop out; any mismatch in the power of $(1-\\alpha)$ or a surviving remainder falsifies the finiteness claim.","supporting_citations":[{"cited_title":"Glimm and A","cited_arxiv_id":null,"evidence_quote":"Supplies the standard framework the paper builds on: analytic continuation of Euclidean fields, the Minlos–Bochner representation of generating functionals, and the P(φ)2 background."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"The author's earlier paper that provides the fractional-Laplacian regularization, the trace-class measure on the finite box, and the determinant expansion (22) on which the cancellation rests."},{"cited_title":"A simple renormalization scheme in random surfa ce theory","cited_arxiv_id":null,"evidence_quote":"Provides the determinant evaluation with the off-diagonal-ones matrix that yields the one-over-square-root-of-(N−1) suppression factor in eq(24)."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Cited for the fact that the massless two-dimensional Green function is not a tempered distribution, which rules out the naive propagator prescription (29)."},{"cited_title":"Botelho, Lecture Notes in Applied Differential Equations o f Mathematical Physics World Scientific, (2008), Singapore ISBN: 10981-281-457 -4","cited_arxiv_id":null,"evidence_quote":"Cited for the Feynman perturbative definition (Chapter 5, §5.2, eq(5.11)) that fixes the order of limits in (18), and for the Wiener–Kac constructions used in Sections 4–5."}],"review_version":1}