{"id":"53279aa5-66c2-4192-bd14-a97598dd35ad","arxiv_id":"2507.03451","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A wavelet transform derivation yields explicit Green-function representations for the Poisson and Helmholtz equations on the n-sphere, with new closed forms for certain negative and noninteger parameters.","lead":"This paper uses continuous spherical wavelets to re-derive the Green functions for the Poisson and Helmholtz equations on the n-dimensional unit sphere and to compute closed-form expressions for some parameter values. The wavelet derivation is new, but the resulting solution formula is the classical spectral Green function, so the value lies mostly in the integral representation and the new closed forms for negative or noninteger parameters.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The Green-function series (16)/(22) is treated as a function, but for n≥5 it is not in L² and at t=1 it diverges pointwise for every n≥2; Theorem 4.3's term-by-term integration therefore lacks a distributional or Sobolev framework.","rationale":"The paper's central contribution is an explicit representation of the Green function as a spectral series and, for special a, as an elementary integral (23), supported by many worked tables. The algebraic derivation from the wavelet construction and the Poisson kernel is coherent, and the integer-L table entries matching [37,38] provide independent support. The weakest point is the convergence framework of the series defining G. Checking the asymptotics confirms the reader's concern and makes it sharper: with coefficients ~1/l and ||C_l^λ||² ~ l^{n-3}, the series is not even in L² for n≥5. The proof of Theorem 4.3 asserts absolute convergence of (24) for r<1 and then integrates term by term; the dominated-convergence hypothesis for the double integral fails, since the sum of the absolute integrated coefficients diverges. The paper's own Remark 4.2(3) concedes that a Sobolev-space treatment is needed and refers to [31]. Hence the identity u=f*G and (23) are formally correct but unproven as stated. This is not a counterexample to the mathematics; a distributional interpretation should rescue the formulas, and the closed forms can be tested directly. The conditional verdict is appropriate, and I would not change it. The concern is a rigor gap in the theorem statements, not a novelty dispute.","tokens_in":16443,"tokens_out":11614,"duration_ms":139661,"concrete_test":"Take the concrete case n=5, L=1/2 (a=9/4) from Table 3. Define the Nth partial sum G_N(cosϑ)=Σ_{l=0}^{N} b_l C_l^λ(cosϑ) with b_l=(λ+l)/(λ)/(a-l(n+l-1)), and compare it with the closed form G(t) given in Table 3 in the distributional sense: for smooth zonal test functions φ(t)=t^k, k=0,...,10, compute ⟨G_N,φ⟩=∫_{-1}^{1}G_N(t)φ(t)(1-t²)^{λ-1/2}dt and the corresponding pairing of G(t). Check whether the difference tends to 0 as N→∞. If it does, (23) is validated as a distributional identity and the gap is presentational; if not, the interchange in Theorem 4.3 is genuinely false. As an auxiliary check, repeat the pairings with an Abel factor e^{-εl}; agreement with the ε→0 limit would confirm Abel summation as the intended interpretation.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The load-bearing step is the identification of the Green kernel as the spectral series G=Σ (λ+l)/(λ) C_l^λ/(a-l(n+l-1)) in (16), (18), (22), and the term-by-term integration in the proof of Theorem 4.3. The paper never specifies the convergence mode. The coefficients behave like 1/l, while the L² normalizations give ||C_l^λ||² ~ l^{n-3}; hence the zonal L² norm of the partial sums behaves as Σ l^{n-5}, diverging for n≥5. Pointwise at t=cosϑ=1 the terms behave like l^{n-3}, so the series diverges for every n≥2 at that point. Thus G is not generally a function; it is a distribution of negative Sobolev order. In Theorem 4.3 the series (24) is declared absolutely convergent for r∈[0,1), but the subsequent interchange with the double integral in (23)-(26) requires dominated convergence, i.e. finiteness of Σ_l |∫∫ r^l dr dR| |(λ+l)/λ C_l^λ|; this fails because the integrated coefficients are ~1/l² but the Gegenbauer factor at t=1 is ~l^{n-3}. The authors themselves flag a related gap in Remark 4.2(3), where they note that [31] treats convergence in spherical Sobolev spaces 'more precisely'. Consequently the central formulas u=f*G and the closed form (23) are, as written, formal identities. They are likely true in a distributional sense, but that framework is not supplied.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops a wavelet-transform method for solving the Poisson and Helmholtz equations on the unit n-sphere. It derives convolution representations u = f * G with G given by a Gegenbauer series (Theorem 4.1) and gives a double-integral closed form for G when a = L(n+L-1) with L non-integer (Theorem 4.3), together with extensive tables of explicit formulas for n = 2,...,10 and various values of a. The method is based on the continuous spherical wavelet transform of the authors' earlier work, and the closed forms are claimed to coincide with known results by Szmytkowski for the positive integer resonance cases.","tokens_in":16786,"tokens_out":5609,"duration_ms":67900,"significance":"If the results are made rigorous, they would provide a new wavelet-based derivation of Green functions on spheres and give closed forms for Helmholtz parameters, in particular negative values of a, that have not been published before. The paper is unusual in offering analytical rather than numerical wavelet solutions, and the tabulated formulas are potentially valuable reference material. However, the manuscript's central analytical objects—the Green function series—are not functions in general, so the stated theorems are formal as written and require a distributional or Sobolev-space interpretation. The paper does supply explicit, verifiable formulas for many cases, which is a strength even where the proofs are incomplete.","major_comments":[{"comment":"The series (16) defining G does not converge pointwise for n >= 2: at t = 1 the terms behave like l^{n-3}, and the zonal L^2 norm of partial sums behaves like sum l^{n-5}, which diverges for n >= 5. For n = 2 the series diverges logarithmically at t = 1. Thus G is not generally an ordinary function, and u = f * G in (15) is not a classical convolution between a continuous function and a function kernel. The proof's claim that 'L2-convergence of the triple integral is ensured' does not establish convergence of the final series (16). The manuscript must state the distributional or Sobolev framework in which (15) holds, and prove the identification of G with that object.","section":"Eq. (16) and the proof of Theorem 4.1"},{"comment":"The interchange of the series (24) with the double integration in (23) is not justified. Although (24) converges absolutely for each r in [0,1), the integrated series is bounded by sum_l ((lambda+l)/lambda) |C_l^lambda(cos theta)| / l^2, which behaves like sum l^{n-3} for n >= 3 and diverges logarithmically for n = 2. Dominated convergence therefore fails. The equality of the double integral with the series should be proved in a distributional sense or by an explicit summation method (such as Abel summation), and this is essential to the validity of the closed-form formula (23).","section":"Proof of Theorem 4.3, Eqs. (24)-(26)"},{"comment":"The paper acknowledges in Remark 4.2(3) that [31] treats convergence of such series in spherical Sobolev spaces 'more precisely,' which indicates the gap is known to the authors. Corollary 4.4 inherits the same issue for integer L: the subtracted series (28) still fails to converge classically for n >= 2. The manuscript should either incorporate a Sobolev-space or distributional framework and prove the convolution statement in that framework, or explicitly state that all Green-function identities are understood distributionally and supply the corresponding convolution theorem for continuous f.","section":"Remark 4.2(3) and Corollary 4.4"}],"minor_comments":[{"comment":"The name 'Klein–Gordan' should be 'Klein–Gordon'.","section":"Introduction, paragraph 2"},{"comment":"The placement of the factor pi in the tabulated expressions is visually ambiguous, because it appears after a large fraction without a clear denominator; please reformat so that the formulas are unambiguous, for example with an explicit multiplication sign or by placing pi in the numerator or denominator consistently.","section":"Tables 3 and 4"},{"comment":"In condition (11), the l = 0 statement is written as an integral equal to zero; it would be clearer to say explicitly that the pair has zero mean, as is done in (12).","section":"Definition 3.1"},{"comment":"The sentence 'If integral u is not 0 and a is not 0, it follows directly from (14)' skips the verification that the l = 0 mode of G is 1/a; adding one line showing that the constant mode is handled correctly would improve clarity.","section":"Proof of Theorem 4.1, last paragraph"},{"comment":"The notation L0 := max{[L], [-n-L+1]} with [x] denoting the integer part is functional but slightly terse; a short parenthetical explanation of why this choice of L0 removes the singular modes would help the reader.","section":"Theorem 4.3"}],"recommendation":"major_revision","confidential_remarks":"The manuscript's self-citations to the authors' prior wavelet papers are reasonable because the method builds directly on that work. The main risk is not novelty but rigor: the closed forms appear to match known results in the tested cases, but the proof of the central identity (23) requires a distributional or Sobolev-space framework. If the authors can supply that framework without changing the substance of the tables, the paper would be acceptable. The fit with math.AP is good, though the paper is also close to approximation theory and special functions."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: the paper is a useful, honest contribution, but it has one real gap — the Green-function series is treated as a function, and the proof of Theorem 4.3 uses a term-by-term integration that the stated bounds don't justify. I think the formulas are right, probably in a distributional or L^p sense, but the paper needs to say so.\n\nWhat's new: the wavelet-transform derivation is new, and Tables 3 and 4 give closed forms for non-integer and negative L that I don't think are in the cited literature. Where the results overlap with Szmytkowski's tables, they match. The paper is also honest about the classical status of the series solution (Remark 4.2(1)) and about the more careful convergence treatment in Michel's book (Remark 4.2(3)).\n\nWhere it's soft: the series (16)/(22) diverges pointwise at t=1 and isn't in L^2 for n≥4. The paper never states the convergence mode. In the proof of Theorem 4.3, the series (24) is absolutely convergent for r<1, but the interchange with the double integral needs dominated convergence; the integrated coefficients decay only like 1/l^2, and with the bound |C_l^λ| ≤ (n+l−2)^{n−2} the sum of absolute values diverges for every n≥2. So the equality (23)=(22) is formal as written. The authors flag a related issue in Remark 4.2(3) for the Poisson 2-sphere case, which is good, but it doesn't cover the general theorem. The fix is to state the identities in an appropriate distributional or Sobolev space, or to justify the interchange by Abel summation or analytic continuation in L.\n\nMinor point: in the proof of Theorem 4.1, equation (2) is cited where the Gegenbauer expansion formula is meant; that's trivial.\n\nMy take: the soft spot is real but repairable, and the new closed forms are useful enough to justify referee time. I'd send it to peer review with a request to add the convergence framework and to compare Tables 3 and 4 more explicitly against [37,38].","headline":"A genuinely new wavelet derivation and useful new closed forms, but the Green-function series needs a distributional framework and the main proof has an unjustified interchange.","tokens_in":17305,"tokens_out":5546,"would_cite":true,"duration_ms":65376,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["42C40","42B37"],"pacs":[],"model":"deepseek-v4-flash","headline":"Wavelet inversion gives explicit Green functions for Poisson and Helmholtz equations on the n-dimensional sphere.","keywords":["spherical wavelets","n-spheres","Poisson equation","Helmholtz equation","Green function","Gegenbauer polynomials","Laplace-Beltrami operator","continuous wavelet transform"],"falsifier":"For $n=2$ and $a=0$, set $t=1$ in the series (16): the $l$-th term is $-\\frac{2l+1}{l(l+1)} \\sim -2/l$, so the series diverges logarithmically and cannot be read as an ordinary pointwise sum. A decisive test is to verify the convolution identity $u = f*G$ in a specified distributional or Sobolev sense for smooth test data with known solutions; if the truncated convolutions do not converge to the exact solution in that norm, the claimed solution formula fails.","tokens_in":16209,"feed_emoji":"🌐","tokens_out":6139,"duration_ms":72925,"temperature":0.7,"pith_summary":"This paper tries to establish that the continuous spherical wavelet transform derived from approximate identities can produce analytical, not just numerical, solutions to the Poisson equation and the Helmholtz equation on the unit sphere in any dimension. The main result is that every classical solution can be written as a convolution of the right-hand side with a zonal Green function whose Gegenbauer coefficients are determined by the eigenvalues of the Laplace-Beltrami operator. For the Poisson equation and for Helmholtz parameters of the special form $a = L(n+L-1)$ with $L \\notin \\mathbb{Z}$, the Green function is rewritten as a double integral involving the Poisson kernel minus a finite polynomial sum, giving closed formulas. The paper also tabulates these closed Green functions for dimensions two through ten and for several positive, negative, and fractional values of $a$. The significance is that exact solution formulas on spheres of arbitrary dimension are rare, and closed forms for negative $a$ appear to be new.","feed_headline":"Wavelet method solves Poisson and Helmholtz on any n-sphere","feed_subtitle":"Explicit Green-function formulas in every dimension, including new closed forms for negative Helmholtz parameters.","key_machinery":"The central object is the zonal Green function $G$, a function of the angular distance on $S^n$ represented by its Gegenbauer expansion; the coefficients $1/(a - l(n+l-1))$ are chosen so that convolution inverts the Laplace-Beltrami operator, whose eigenvalues are $-l(n+l-1)$. The argument is carried by the continuous spherical wavelet transform based on approximate identities: applying $\\Delta_* + a$ to the analyzing Poisson wavelet produces an admissible wavelet family, and the reconstruction formula converts the differential equation into a convolution with $G$. The closed-form part relies on the Poisson kernel identity $\\Sigma_n p_r(\\cos\\theta) = \\sum_{l=0}^\\infty r^l \\frac{\\lambda+l}{\\lambda} C_l^\\lambda(\\cos\\theta)$ together with the integral identity $\\int_0^1 R^{-(n+2L)} \\int_0^R r^{n+L-2} r^l\\, dr\\, dR = -\\frac{1}{(L-l)(n+L+l-1)}$, which reproduces the desired coefficient exactly.","core_discovery":"The paper claims that if $f \\in C(S^n)$ and $u \\in C^2(S^n)$ solve $\\Delta_* u + a u = f$ with $a$ not equal to any eigenvalue $l(n+l-1)$, then $u = f * G$, where $G$ is the zonal series $G = \\sum_{l=0}^\\infty \\frac{1}{a - l(n+l-1)} \\frac{\\lambda+l}{\\lambda} C_l^\\lambda$, with $\\lambda = (n-1)/2$ and $C_l^\\lambda$ the Gegenbauer polynomials. For the Poisson equation with zero-mean data the same formula holds starting at $l=1$ with coefficient $-1/(l(n+l-1))$. For $a = L(n+L-1)$ with non-integer $L$, Theorem 4.3 expresses $G$ as a double integral of the Poisson kernel minus the first $L_0$ Gegenbauer terms, plus a finite rational correction; this is the mechanism that yields closed forms. The paper states that for tested cases these closed forms coincide with known Green functions, including the classical logarithmic kernel on $S^2$, and that the negative-$a$ cases have not appeared in closed form before.","pith_inferences":["If the distributional interpretation of the Green series is made explicit and rigorous, the same wavelet-inversion trick should produce Green functions for other rotation-invariant operators on $S^n$, such as polyharmonic operators or fractional powers of $\\Delta_*$.","The double-integral representation in Theorem 4.3 suggests a quadrature-based numerical evaluation that avoids summing the divergent Gegenbauer series, which would be stable away from the singular diagonal.","Since the integral representation is valid for all non-integer $L$, the tabulated closed forms could be extended systematically to arbitrary rational $a$ by symbolic integration.","Because the kernel is zonal, the method should extend to non-zonal right-hand sides through the paper's convolution framework, which already handles general $L^2$ functions by rotation-invariant kernels."],"forward_implications":["For every dimension $n \\ge 2$ and every $a$ that is not an eigenvalue, the Helmholtz solution on $S^n$ is given by a single convolution formula, so the problem is reduced to evaluating one zonal kernel.","For $a=0$ the Poisson Green function has explicit closed forms in dimensions $2$ through $10$, including the standard logarithmic kernel on the two-sphere.","For Helmholtz parameters $a = L(n+L-1)$ with integer $L$, the formula with the critical term removed gives the generalized Green function and reproduces previously known closed forms.","For negative $a$, such as $a = -3/4$ on $S^3$ or $a = -2$ on $S^4$, the closed forms in Table 4 are presented as new, with applications to telegraph and Klein-Gordon type equations.","The derivation provides an analytical alternative to numerical wavelet solvers, producing exact series and closed integrals rather than approximate solutions."],"supporting_citations":[{"why":"Gives the closed-form Helmholtz Green function on the two-dimensional unit sphere that the paper's Table 2 entries are checked against.","marker":"[37]"},{"why":"Provides closed forms of the generalized Green function for the Helmholtz operator on the $N$-dimensional sphere, used to verify the paper's higher-dimensional formulas.","marker":"[38]"},{"why":"Supplies the Poisson-equation Green function on the two-sphere that the $n=2$, $a=0$ result is compared with.","marker":"[13]"},{"why":"Gives an alternative standard form of the two-sphere Poisson Green function, which the paper notes its result matches up to normalization.","marker":"[12]"},{"why":"Supplies the convergence proof for the simplified wavelet transform that the paper adapts to its setting.","marker":"[22]"},{"why":"Provides the eigenvalue identity $\\Delta_* Y_l^k = -l(n+l-1) Y_l^k$ for hyperspherical harmonics, used to invert the operator spectrally.","marker":"[35]"},{"why":"Gives the uniform bound $|C_l^\\lambda(\\cos\\vartheta)| \\le (n+l-2)^{n-2}$ that justifies absolute convergence and termwise integration in Theorem 4.3.","marker":"[36]"}],"fun_headline_variants":["Wavelet method yields explicit Green functions on any n-sphere","Closed-form Green functions for Poisson and Helmholtz on n-spheres","Explicit Poisson and Helmholtz solutions on any n-sphere","New closed forms for spherical Helmholtz and Poisson equations","Wavelet method solves spherical Poisson and Helmholtz"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The formulas work only if the formal Gegenbauer series for $G$ is given a meaning beyond pointwise convergence, because at the north pole the series diverges logarithmically even in the classical two-dimensional Poisson case.","fun_headline_variants_meta":{"raw":{"variants":["Wavelet method yields explicit Green functions on any n-sphere","Closed-form Green functions for Poisson and Helmholtz on n-spheres","Explicit Poisson and Helmholtz solutions on any n-sphere","New closed forms for spherical Helmholtz and Poisson equations","Wavelet method solves spherical Poisson and Helmholtz"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001051,"raw_usage":{"total_tokens":4369,"prompt_tokens":857,"completion_tokens":3512,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":473,"completion_tokens_details":{"reasoning_tokens":3430}},"tokens_in":473,"tokens_out":3512,"duration_ms":31100,"temperature":1.0,"reasoning_tokens":3430,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T20:11:45.666694+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"For $n=2$ and $a=0$, set $t=1$ in the series (16): the $l$-th term is $-\\frac{2l+1}{l(l+1)} \\sim -2/l$, so the series diverges logarithmically and cannot be read as an ordinary pointwise sum. A decisive test is to verify the convolution identity $u = f*G$ in a specified distributional or Sobolev sense for smooth test data with known solutions; if the truncated convolutions do not converge to the exact solution in that norm, the claimed solution formula fails.","supporting_citations":[{"cited_title":"Szmytkowski, Closed form of the generalized Green ’s function for the Helmholtz operator on the two-dimensional unit sphere , J","cited_arxiv_id":null,"evidence_quote":"Gives the closed-form Helmholtz Green function on the two-dimensional unit sphere that the paper's Table 2 entries are checked against."},{"cited_title":"Szmytkowski, Closed forms of the Green ’s function and the generalized Green ’s function for the Helmholtz operator on the N -dimensional unit sphere , J","cited_arxiv_id":null,"evidence_quote":"Provides closed forms of the generalized Green function for the Helmholtz operator on the $N$-dimensional sphere, used to verify the paper's higher-dimensional formulas."},{"cited_title":"Freeden, M","cited_arxiv_id":null,"evidence_quote":"Supplies the Poisson-equation Green function on the two-sphere that the $n=2$, $a=0$ result is compared with."},{"cited_title":"Freeden, M","cited_arxiv_id":null,"evidence_quote":"Gives an alternative standard form of the two-sphere Poisson Green function, which the paper notes its result matches up to normalization."},{"cited_title":"Iglewska-Nowak, A continuous spherical wavelet transform for C(Rn), Appl","cited_arxiv_id":null,"evidence_quote":"Supplies the convergence proof for the simplified wavelet transform that the paper adapts to its setting."},{"cited_title":"Shimakura, Partial differential operators of elliptic type , Translations of Mathematical Mono- graphs, Vol","cited_arxiv_id":null,"evidence_quote":"Provides the eigenvalue identity $\\Delta_* Y_l^k = -l(n+l-1) Y_l^k$ for hyperspherical harmonics, used to invert the operator spectrally."},{"cited_title":"Szeg¨ o,Orthogonal polynomials, Fourth Edition, AMS Coll","cited_arxiv_id":null,"evidence_quote":"Gives the uniform bound $|C_l^\\lambda(\\cos\\vartheta)| \\le (n+l-2)^{n-2}$ that justifies absolute convergence and termwise integration in Theorem 4.3."}],"review_version":1}