{"id":"313c32c4-75e6-453e-8e6e-7a44f54b2557","arxiv_id":"1908.02161","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For the operator (−∆)^ν, the heat kernel is a generalized exponential (Fox-Wright) function E_{ν,d/2}(−x^2/4τ^{1/ν}), with power-law asymptotics for noninteger ν and oscillatory exponential asymptotics for integer ν.","lead":"The paper derives the heat kernel of the ν-th power of the Laplacian using a generalized exponential function, showing that its short-time asymptotics oscillate and, for integer orders, match a resummed semiclassical expansion that fails in the coincidence limit. It provides a mathematical building block for heat-kernel calculations in higher-derivative and nonlocal quantum gravity models.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Full integer-ν asymptotics (4.20) rests on the unproven gamma-ratio expansion (A.7); leading-order oscillatory decay is independently supported, but the complete series needs explicit E_m and validity bounds.","rationale":"The paper's main constructive result is the exact representation (1.3) with GEF; that part is well supported: the Taylor coefficient matching in Sec. 2 is valid because of homogeneity and entireness for ν>1/2, and the Mellin–Barnes machinery is standard. The asymptotic classification for noninteger ν (power-law) follows from (3.3) and is not in dispute. The load-bearing point is the integer-ν asymptotic expansion (4.20), because it is used both to claim the oscillatory exponential form and to diagnose nonuniformity of WKB. That expansion depends crucially on (A.7), which is quoted from H-function theory and only formally sketched. The absence of explicit E_m and remainder bounds means a reader cannot verify the central asymptotic formula; this is exactly the kind of omitted support that warrants a conditional verdict. However, the leading order of (4.22) is independently derived by steepest descent in Sec. 4.2, and the nonuniformity argument only requires the leading saddle structure. So the concern does not overturn the qualitative central claims; it leaves the complete expansion conditional. The deferred companion paper [71] supports withholding ACCEPT.","tokens_in":23325,"tokens_out":12794,"duration_ms":138540,"concrete_test":"Compute E_0, E_1, E_2 for the case N=2, d=1 (equivalently ν=2, α=1/2) explicitly from the Appendix B prescription (D_n, C_n, inverse matrix d^{-1}_{kj}) and evaluate the truncated RHS of (4.16) against the exact ratio ε̃_{2,1/2}(s) at, say, s=1+it for t=20,50,100 (and also s=2+it). If the relative error does not decrease like |s|^{-M-1} as the truncation order M and |s| grow, expansion (A.7) is not an asymptotic expansion in the needed sector. Then check the corresponding truncated (4.20) against direct numerical evaluation of K_{2,1}(τ,x) from (2.2) for τ=10^{-4} and x=0.5; agreement at the expected order would confirm the termwise Mellin inversion.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central quantitative claim for integer powers is Eq. (4.20), the complete asymptotic expansion of the heat kernel. Its derivation goes through the Mellin–Barnes representation (3.1), the sine decomposition (4.8)–(4.11), and the key step (4.16): the expansion (A.7) of the gamma-function ratio in (4.9), followed by termwise inverse Mellin transform (4.13)→(4.17) using (4.6). The paper states that E_m are 'systematically calculable' (Appendix B) but gives no explicit E_m, no sector of validity in arg s, and no remainder estimate on the vertical contour C_w. Appendix B is a formal Stirling manipulation: coefficients D_n and C_n are defined, an infinite lower-triangular inversion (B.6)–(B.8) is asserted, but no proof that the resulting double series yields a Poincaré asymptotic expansion uniform in s on C_w is provided. If (A.7) is only a formal expansion or fails in the relevant sector, termwise Mellin inversion is unjustified and the correction terms in (4.20) could be wrong, weakening the claimed WKB comparison beyond leading order. The leading m=0 term is independently reproduced by steepest descent (Sec. 4.2), so the qualitative oscillatory exponential behavior and nonuniformity at x=0 are robust; the unresolved part is the full asymptotic series and its coefficients. The promised Schwinger–DeWitt application is deferred to [71], so the physical reach of the paper is not yet demonstrated.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the heat kernel K_{ν,d}(τ,x) = e^{-τ(-Δ)^ν}δ(x) in flat d-dimensional Euclidean space. It derives the exact representation K_{ν,d}(τ,x) = (4πτ^{1/ν})^{-d/2} E_{ν,d/2}(-x^2/(4τ^{1/ν})), where E_{ν,α}(z) is a two-parameter generalized exponential function defined by a Taylor series of gamma-function ratios. The function is identified with a Fox–Wright Ψ-function, and several representations are given: a Mellin–Barnes integral, a Bessel–Clifford integral, and a closed form for ν=1/2. The paper then studies large-z asymptotics, finding power-law falloff for noninteger ν and oscillatory exponential behavior for integer ν. For integer N, a complete asymptotic expansion is claimed via the Mellin–Barnes representation and the theory of Fox H-functions, and it is compared with the Pauli–Van Vleck/WKB ansatz and with steepest descent. The paper concludes that the WKB heat-kernel expansion is not uniform in the coincidence limit x=0 for ν>1, and announces upcoming applications to higher-derivative and Horava–Lifshitz-type operators.","tokens_in":23625,"tokens_out":7573,"duration_ms":88006,"significance":"If the main results hold, the paper provides a useful explicit building block for heat kernels of higher-derivative and nonlocal operators, and it correctly identifies why a naive WKB ansatz cannot be used for such operators. The exact formula (1.3), the Mellin–Barnes representation, the ν=1/2 closed form, and the exact normalization check in Sec. 4.3 are cleanly derived and appear trustworthy. The leading-order integer-N asymptotics are independently confirmed by steepest descent in Sec. 4.2, which makes the main qualitative conclusion robust. However, the claimed complete asymptotic expansion for integer N depends on an imported gamma-ratio expansion whose proof, validity sector, coefficients, and remainder estimates are not given in the manuscript. The paper therefore is reliable in its leading-order physics but does not yet fully establish its central all-orders asymptotic claims.","major_comments":[{"comment":"The definition of E_{ν,α}(z) by the Taylor series (1.4) is problematic for ν<1/2, because, as the paper itself states in Sec. 3, that series diverges and is only asymptotic for z→0 in this regime. The derivation of Eq. (1.3) in Sec. 2 from equality of derivatives at x=0 is therefore incomplete for such ν; equality of formal Taylor coefficients does not determine a non-analytic function. The authors should define E_{ν,α}(z) globally via the Mellin–Barnes representation (3.1) or the Bessel–Clifford integral (2.16), and state explicitly that (1.4) is the convergent Taylor expansion for ν>1/2 but only a formal or asymptotic series for ν<1/2. Since Eq. (1.3) is asserted for generic ν, this clarification is needed to make the main exact statement precise.","section":"§4.1 and Appendix A/B, Eqs. (4.16)–(4.20)"},{"comment":"The argument that the WKB expansion fails to reproduce the initial condition is weakened by an unjustified termwise integration of an asymptotic expansion. Eq. (4.32) sums all N branches K^(j), but footnote 3 acknowledges that branches with j≠0,N−1 are exponentially subdominant and should be discarded; moreover, an asymptotic expansion cannot be integrated termwise near x=0 without uniformity control. The nonuniformity conclusion is nevertheless correct and can be made cleanly: the exact value (4.34) is finite and nonzero at x=0, while the asymptotic expansion (4.22) is singular at x=0 for N>1. Please restate Sec. 4.3 so that the logical gap in the integration argument is removed and the nonuniformity claim rests on the direct comparison of (4.22) with (4.34).","section":"§4.3, Eqs. (4.32)–(4.34)"}],"minor_comments":[{"comment":"The expression (n/ν)! in Eq. (3.10) is only meaningful when n/ν is a nonnegative integer; please state this condition explicitly before using the notation.","section":"Eq. (3.10)"},{"comment":"The symbol d^β/dz^β in Eq. (3.11) is not defined for noninteger β; if a fractional integro-differentiation operator is intended, specify the convention, or restrict the identity to integer β with a remark about the fractional extension.","section":"Eq. (3.11)"},{"comment":"The switch between ν and N in Sec. 4 is introduced explicitly, but a short remark in the text or a footnote clarifying that N denotes positive integers while ν is generic would improve readability.","section":"Sec. 4"},{"comment":"The figures are informative, but the overlapping curves in Fig. 3 would be easier to read with additional line styles or markers, and the figure captions could state the value of the fixed parameters more prominently.","section":"Figs. 2–4"}],"recommendation":"major_revision","confidential_remarks":"For the editor: the main obstacle to acceptance is the unsupported all-orders asymptotic expansion (A.7) used in Sec. 4.1. The leading-order result is independently confirmed, so the paper is not fatally flawed, but the central quantitative claim needs either a rigorous statement of the H-function asymptotic theorem with explicit conditions and coefficients, or a clear reduction of the claim to leading order. The authors' announced follow-up paper [71] should not be used as a substitute for completing this argument. The manuscript is otherwise within the scope of the journal and the topic is likely of interest to the readership."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this paper is worth engaging. The central representation (1.3) is not new — Lee-Pac, Gusynin, and the anomalous diffusion literature already had it — but the paper does two genuinely useful things. It gives a systematic asymptotic classification of the GEF, separating power-law decay for noninteger ν from oscillatory exponential behavior for integer ν, and it spells out a point that has been underappreciated: the semiclassical WKB ansatz for higher-order heat kernels is not uniform in the coincidence limit for ν>1. That last point, with the explicit comparison in Sec. 4.3, is the most valuable takeaway for anyone trying to build a Schwinger-DeWitt machinery for higher-derivative gravity.\n\nThe math is mostly clean. The derivation of (1.3) from the momentum integral is straightforward, the Mellin-Barnes representation and the ν=1/2 closed form (3.7) are correct, and the steepest descent analysis in Sec. 4.2 independently reproduces the leading term of the integer-ν expansion. The connection to Fox-Wright and H-functions is well placed.\n\nThe soft spot is Appendix A.7. The full asymptotic series (4.20) rests on the gamma-ratio expansion (A.7), which is quoted from Braaksma with coefficients E_m described only as 'systematically calculable.' Appendix B is a formal Stirling manipulation; it asserts the inversion (B.8) but does not prove convergence, validity sector in arg s, or remainder bounds. That is a real gap, but it is not fatal. The leading m=0 term is independently validated by steepest descent, so the qualitative results — exponential oscillatory behavior, the power-law vs exponential distinction, and the nonuniformity at x=0 — are solid. What is unresolved is whether the subleading corrections in (4.20) are exactly right. I would not build a precision calculation on those coefficients until the expansion is made explicit.\n\nThe citation pattern is honest. The authors credit the earlier appearances of the representation, and the self-citations are context. The promised recurrence relations are deferred to [71], so the physical payoff is not yet demonstrated. Minor novelty inflation in the intro, but they walk it back.\n\nRecommendation: the paper deserves a serious referee, not a desk reject. The referee should ask for a more solid treatment of (A.7) — either explicit E_m and validity bounds, or a clear statement that the full series is heuristic beyond leading order. It is a good candidate for publication after that.","headline":"Solid paper: restates a known representation but adds systematic asymptotics and a clear warning about WKB nonuniformity; the main gap is the unproven gamma-ratio expansion behind the full integer-ν series.","tokens_in":24215,"tokens_out":2123,"would_cite":true,"duration_ms":22217,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["58J35","35K08","41A60"],"pacs":[],"model":"deepseek-v4-flash","headline":"The heat kernel of (−Δ)^ν is exactly a generalized exponential function, and its asymptotics split sharply between oscillatory-exponential integer order and power-law fractional order.","keywords":["heat kernel","fractional Laplacian","higher-derivative operators","generalized exponential function","Fox-Wright function","Fox H-function","proper-time expansion","oscillatory heat kernel"],"falsifier":"Evaluate the momentum integral (2.2) numerically to high precision for $N=3$, $d=4$, $\\tau=1$ over a range of $x$, compare it with the truncated exact series (1.4), and test whether the leading two-branch asymptotic formula (4.22) reproduces the oscillation period, amplitude, and the prefactor $x^{-d(N-1)/(2N-1)}$; a mismatch in any of these would falsify the central asymptotic claim.","tokens_in":23106,"feed_emoji":"📉","tokens_out":13440,"duration_ms":135093,"temperature":0.7,"pith_summary":"This paper establishes the exact heat kernel of the operator $(-\\Delta)^{\\nu}$ in flat $d$-dimensional Euclidean space: up to a universal prefactor it is the generalized exponential function $E_{\\nu,d/2}(-x^2/4\\tau^{1/\\nu})$, defined by a single Taylor series. The reason this matters is that the usual exponential (WKB) ansatz for the Laplacian cannot be extended to higher powers, because the short-proper-time expansion contains infinitely many negative powers of $\\tau$; the generalized exponential resums all of them. The paper derives the kernel's large-distance and small-$\\tau$ behavior and finds a sharp split: noninteger $\\nu$ gives power-law falloff, while integer order $N$ gives a superposition of $N$ oscillatory exponential branches. It also shows that the semiclassical expansion for $N>1$ is not uniform in the coincidence limit $x=0$, which constrains any attempt to build a proper-time heat-kernel expansion for higher-derivative quantum field theories.","feed_headline":"Heat kernels of (-Δ)^ν are generalized exponentials, not Gaussians","feed_subtitle":"Exact formula resums all negative proper-time powers; integer orders oscillate, fractional orders decay as power laws.","key_machinery":"The central object is the generalized exponential function $E_{\\nu,\\alpha}(z)$ of Eq. (1.4), a two-parameter entire function for $\\nu>1/2$ that reduces to $\\exp(z)$ at $\\nu=1$ and to a Bessel–Clifford function as $\\nu\\to\\infty$. Its Mellin–Barnes integral (3.1) turns the heat kernel problem into the theory of Fox–Wright $\\Psi$- and Fox $H$-functions: the ratio of gamma functions in the integrand is expanded by Eq. (A.7), and the inverse Mellin transform (4.6) converts each term into an exponential. For integer $N$, the sine-factor decomposition (4.8)–(4.11) splits the kernel into $N$ 'second-kind' generalized exponentials whose phases are exactly the fractional-power branches of the Hamilton–Jacobi action, which is the mechanism that produces the oscillatory exponential asymptotics.","core_discovery":"For $F=(-\\Delta)^{\\nu}$ in flat space the heat kernel is exactly $$K_{\\nu,d}(\\tau,x)=\\frac{1}{(4\\pi\\$tau^{{1/\\nu}}$)^{d/2}}\\,E_{\\nu,d/2}\\!\\left(-\\frac{$x^{2}$}{4\\$tau^{{1/\\nu}}$}\\right),\\qquad E_{\\nu,\\$\\alpha$}(z)=\\frac{1}{\\nu}\\sum_{m=0}^{\\infty}\\frac{\\Gamma((\\$\\alpha$+m)/\\nu)}{\\Gamma(\\$\\alpha$+m)}\\frac{z^m}{m!}.$$ The function $E_{\\nu,\\alpha}$ is a Fox–Wright $\\Psi$-function, so its Mellin–Barnes representation controls all asymptotics. For noninteger $\\nu$ the large-$z$ (small-$\\tau$ or large-$\\lvert x\\rvert$) limit is a power series in $z^{-\\nu}$, while for integer $N$ all residues cancel and the kernel becomes a sum of $N$ exponential branches with phases $\\phi_j=\\pi(1-N+2j)/(2N-1)$. The two complex-conjugate dominant branches reproduce the semiclassical Pauli–Van Vleck amplitude with a definite choice of phase, and the same result follows from steepest descent. The expansion is shown to be nonuniform at $x=0$ for $N>1$, so the coincidence limit must be taken from the exact GEF rather than from its asymptotics.","pith_inferences":["A natural consequence the authors leave implicit is that UV calculations in higher-derivative gravity should work with split-point GEF kernels rather than only with coincidence-limit heat-kernel coefficients, since the nonuniformity means the diagonal expansion cannot see the full short-distance structure.","The integer-versus-fractional dichotomy suggests a concrete diagnostic: local higher-derivative propagators should show damped oscillations at large separation, while nonlocal noninteger-order propagators should show algebraic tails; numerical studies of such propagators could test this directly.","The exact closed form at $\\nu=1/2$ hints that other rational values $\\nu=p/q$ may also reduce to known special functions, and working these out would give explicit exact heat kernels that independently check the Fox-H asymptotic machinery.","The paper notes that a uniform asymptotic expansion valid across $x\\to0$ is open; finding one would cure the coincidence-limit problem and is a testable mathematical extension."],"forward_implications":["For a local higher-derivative operator of order $2N$, the short-proper-time expansion runs in powers $\\tau^{j/N}$ with coefficients built from GEF values at zero, not from the nonuniform WKB expansion.","For nonlocal operators $(-\\Delta)^{\\nu}$ with noninteger $\\nu$, the heat kernel decays as a power law at large separation, so the standard semiclassical $\\hbar$-expansion does not apply to these kernels.","The exact GEF representation provides the building block for the curved-space expansion announced by the authors, in which generalized heat-kernel coefficients obey recurrent equations.","The heat kernel of $\\sqrt{-\\Delta}$ in flat space is exactly the power-law kernel of Eq. (3.8), which is the massless limit of a simple brane-to-bulk propagator."],"supporting_citations":[{"why":"It first used the functions (1.3) as heat kernels in higher-derivative models, the starting point that the new resummation extends.","marker":"[38]"},{"why":"It supplied the series expansion (1.4) and earlier heat-kernel asymptotics for nonminimal higher-derivative operators.","marker":"[45]"},{"why":"Wright's asymptotic expansions of generalized hypergeometric functions underpin the Fox–Wright treatment used here.","marker":"[51, 52]"},{"why":"This work's Barnes-integral asymptotics is the source of Eq. (A.7), from which the integer-order exponential branch sum is derived.","marker":"[53]"},{"why":"It gives heat-kernel asymptotics for roots of generalized Laplacians by zeta-function methods, the comparison point for the paper's split-point result.","marker":"[25]"},{"why":"It provides the generalized proper-time technique that this paper extends, and its split-point ansatz is the context for the GEF replacement.","marker":"[6]"},{"why":"It supplies the steepest-descent contour existence and saddle-point selection that confirm the branch phases independently.","marker":"[70]"}],"fun_headline_variants":["Heat kernels of fractional Laplacians are Fox-Wright functions","Integer-order heat kernels oscillate, fractional ones decay as power laws","Exact heat kernel for (-Δ)^ν: no Gaussian, but generalized exponentials","Heat kernel of Laplacian powers: oscillatory branches for integer ν","Non-Gaussian heat kernel: from Fox-Wright to oscillatory sums"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The asymptotic formulas for the integer-order kernel rest on an asymptotic expansion for ratios of gamma-function products quoted from Fox H-function theory, whose coefficients the paper only describes as systematically calculable and whose region of validity is not stated; if that expansion fails or is nonuniform in the sector where it is used, the claimed exponential and oscillatory asymptotics would not follow.","fun_headline_variants_meta":{"raw":{"variants":["Heat kernels of fractional Laplacians are Fox-Wright functions","Integer-order heat kernels oscillate, fractional ones decay as power laws","Exact heat kernel for (-Δ)^ν: no Gaussian, but generalized exponentials","Heat kernel of Laplacian powers: oscillatory branches for integer ν","Non-Gaussian heat kernel: from Fox-Wright to oscillatory sums"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000648,"raw_usage":{"total_tokens":3053,"prompt_tokens":1102,"completion_tokens":1951,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":718,"completion_tokens_details":{"reasoning_tokens":1855}},"tokens_in":718,"tokens_out":1951,"duration_ms":12507,"temperature":1.0,"reasoning_tokens":1855,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T14:52:07.950581+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Evaluate the momentum integral (2.2) numerically to high precision for $N=3$, $d=4$, $\\tau=1$ over a range of $x$, compare it with the truncated exact series (1.4), and test whether the leading two-branch asymptotic formula (4.22) reproduces the oscillation period, amplitude, and the prefactor $x^{-d(N-1)/(2N-1)}$; a mismatch in any of these would falsify the central asymptotic claim.","supporting_citations":[{"cited_title":"Heat equation asymptotics of elliptic operators with non-scalar leading symbol,","cited_arxiv_id":null,"evidence_quote":"It first used the functions (1.3) as heat kernels in higher-derivative models, the starting point that the new resummation extends."},{"cited_title":"Higher-derivative op- erators and DeWitts WKB ansatz,","cited_arxiv_id":null,"evidence_quote":"It supplied the series expansion (1.4) and earlier heat-kernel asymptotics for nonminimal higher-derivative operators."},{"cited_title":"Local heat ker- nel asymptotics for nonminimal differential operators,","cited_arxiv_id":null,"evidence_quote":"This work's Barnes-integral asymptotics is the source of Eq. (A.7), from which the integer-order exponential branch sum is derived."},{"cited_title":"The spectral geometry of a Riemannian manifold,","cited_arxiv_id":null,"evidence_quote":"It gives heat-kernel asymptotics for roots of generalized Laplacians by zeta-function methods, the comparison point for the paper's split-point result."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"It supplies the steepest-descent contour existence and saddle-point selection that confirm the branch phases independently."}],"review_version":1}