{"id":"a4ce4fad-d587-4063-b49c-a40a30bcab0e","arxiv_id":"1908.08561","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The paper re-derives the 2012 second-order perturbative spectral sum rules of rational order via a new trace and fractional Green's function construction.","lead":"A mathematical physics paper presents a trace-based method, using fractional Green's functions, to compute rational-order sum rules for quantum billiards with nonuniform density. It reproduces a second-order perturbative result previously obtained by the same author in 2012, now derived without Rayleigh-Schrödinger perturbation theory.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Eq. (37)'s off-diagonal term has a wrong exponent: the derivation gives epsilon_m^{1-s}, not epsilon_m^{-s}; as printed the central sum rule fails a two-level perturbative check.","rationale":"The reader identified perturbative convergence and term-by-term trace interchange as the weakest assumption. My independent check of the algebra leading from eqs. (32)-(35) to eqs. (36)-(37) found a more immediate, internally checkable problem: the off-diagonal exponent in the final sum rule is printed as epsilon_m^{-s} but the derivation yields epsilon_m^{1-s}. A two-level calculation settles the issue without any infinite-series interchange, and it shows the printed eq. (37) is not the correct second-order coefficient, while the corrected version is. This supports a CONDITIONAL verdict rather than rejection: the trace-based method appears coherent, the errors are localized typos in displayed formulas, and the agreement with Ref. [1] should hold after the exponent correction. I also note the typo in eq. (3), where sqrt(Sigma(x)) appears twice and the second factor should be sqrt(Sigma(y)). These issues reinforce the reader's conditional stance, so I recommend no change to the verdict, but the author should correct eqs. (36)-(37), fix eq. (3), and confirm the corrected eq. (37) against eq. (9) of Ref. [1].","tokens_in":9027,"tokens_out":40523,"duration_ms":367068,"concrete_test":"Use a two-dimensional truncation with H0=diag(1,4), S=I+lambda t(|1><2|+|2><1|), and s=3/2. Expand the exact eigenvalues of S^{-1/2} H0 S^{-1/2} to second order in lambda, or diagonalize numerically for a small fixed t and subtract the unperturbed value, to obtain the lambda^2 coefficient of Z(s). Compare this coefficient with (i) eq. (37) as printed and (ii) eq. (37) with epsilon_m^{-s} replaced by epsilon_m^{1-s}. If only (ii) matches, correct the exponent in eqs. (36)-(37) and re-verify the stated agreement with eq. (9) of Ref. [1].","verdict_should_be":"UNCHANGED","load_bearing_attack":"The algebra leading from eq. (34) to eqs. (36)-(37) does not produce the displayed off-diagonal factor. The intermediate result is Z^{(2)}(s) = (lambda^2/2)s [ (s-1) sum_n <sigma>^2/epsilon_n^s + T ], with T = sum_{n != m} (epsilon_n^{1-s} - epsilon_m^{1-s})/(epsilon_m - epsilon_n) |<n|sigma|m>|^2 = - sum_{n != m} (epsilon_m^{1-s} - epsilon_n^{1-s})/(epsilon_m - epsilon_n) |<n|sigma|m>|^2. Therefore the final formula should contain (epsilon_m^{1-s} - epsilon_n^{1-s})/(epsilon_m - epsilon_n), not (epsilon_m^{-s} - epsilon_n^{1-s})/(epsilon_m - epsilon_n). The printed '-s' is not cosmetic: for the exactly solvable two-level model epsilon_1=1, epsilon_2=4, sigma=t(|1><2|+|2><1|), s=3/2, expanding the eigenvalues of S^{-1/2}(-Delta)S^{-1/2} to second order in lambda gives a lambda^2 coefficient t^2/4 for Z(s). Eq. (37) as printed gives 11 t^2 / 32; replacing epsilon_m^{-s} by epsilon_m^{1-s} gives t^2/4. Since the paper claims eq. (37) reproduces eq. (9) of Ref. [1], the printed formula cannot be the reproduced result. The trace-based derivation itself supports the corrected exponent, so the error is localized but load-bearing for the paper's central display equation.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a trace-based route to spectral sum rules Z(s)=Σ_n E_n^{-s} of rational order for the Helmholtz equation with positive density, -Δψ = E Σ(x)ψ, on a domain where the homogeneous eigenfunctions are known. The author introduces Green's functions of order 1/N, expands their spectral coefficients perturbatively in the density contrast Σ=1+λσ, and expresses Z(1+1/N) and Z(1/N+1/N') as traces of products of these Green's functions. The central displayed result is the second-order formula (37), which is claimed to reproduce Eq. (9) of the author's earlier paper (Ref. [1]). The paper also reports higher-order coefficients for the Green's function of order 1/2 up to eighth order in an appendix.","tokens_in":9236,"tokens_out":38962,"duration_ms":330514,"significance":"If correct, the construction provides an alternative derivation of the known second-order density expansion of spectral zeta functions for inhomogeneous systems and offers a systematic route to higher orders that avoids degenerate Rayleigh-Schrödinger perturbation theory. The Green's-function-of-fractional-order formalism is a sensible and potentially useful idea, and Eq. (32) is internally consistent with a direct perturbative computation. However, the central display equation contains an exponent error: Eq. (37) as printed fails an elementary two-level check, so the claimed agreement with Ref. [1] is not established by the manuscript. The error is localized and correctable, but it is load-bearing for the paper's main claim, and the absence of any independent validation allowed it to pass.","major_comments":[{"comment":"The off-diagonal term in Eq. (36) and hence in Eq. (37) has the wrong exponent. Starting from Eqs. (34) and (35), the correct reduction is Z^{(2)}(s) = (λ²/2)s [ (s-1)Σ_n <n|σ|n>²/ε_n^s + Σ_{n≠m} (ε_n^{1-s}-ε_m^{1-s})/(ε_m-ε_n) |<n|σ|m>|² ], equivalently with a minus sign and the numerator (ε_m^{1-s}-ε_n^{1-s})/(ε_m-ε_n). The printed Eq. (36) instead contains ε_m^{-s} - ε_n^{1-s}. For the two-level model ε1=1, ε2=4, σ=t(|1><2|+|2><1|), and s=3/2, expanding the eigenvalues of S^{-1/2}(-Δ)S^{-1/2} to second order gives the λ² coefficient t²/4 for Z(s); the printed Eq. (37) gives 11t²/32, whereas the corrected exponent gives t²/4. Since Eq. (37) is the paper's central result, this is a load-bearing error, even though it is local and fixable.","section":""},{"comment":"The order-by-order trace expansion is performed by inserting the λ-expansions of q_nm and Q_nm into infinite sums over the spectral basis and rearranging the series. The paper does not justify the interchange of the λ-expansion with the infinite sums, nor does it prove convergence of the perturbation series for q_nm and Q_nm. The assumption |σ(x)|<<1 makes the expansion plausible, but without a convergence or asymptotic justification the word 'exact' in the title and abstract overstates what is demonstrated. The authors should state explicitly that the calculation is order-by-order in λ and provide either a finite-dimensional truncation argument or an analytic-continuation justification for the trace identities.","section":""},{"comment":"The only validation of the final formula is the statement that Eq. (37) agrees with Eq. (9) of Ref. [1]. This is a self-citation and, as the two-level check shows, it is not sufficient: the printed formula does not reproduce the exact second-order coefficient. The paper would be substantially stronger if it included at least one independent check, such as the exactly solvable two-level model above, a finite-dimensional matrix diagonalization, or a numerical solution of a one-dimensional inhomogeneous string. The absence of such a check should be addressed in revision.","section":""}],"minor_comments":[{"comment":"The summation indices in Eq. (10) are incorrect as written: the product Σ_{n,m,m'} q_{nm} q_{n'm'} ψ_n ψ_{m'} does not make sense; the second factor should be q_{m m'} (with an implied sum over the intermediate index).","section":""},{"comment":"The section is titled 'Order 1/N + 1/N'' but the displayed expression for Z begins with Z(1+1/N); it should read Z(1/N+1/N'). This appears to be a typographical error.","section":""},{"comment":"In the zeroth-order coefficient Q^{(0)}_{nm} = δ_{nm}/ε_r, the index r is undefined; the denominator should be ε_n.","section":""},{"comment":"The expression for q^{(8)} contains a double plus sign '++' in the double-sum term; this is a typographical issue.","section":""},{"comment":"The notation q[1/N](k)_{rn} is used before it is defined; the superscript (k) denoting the order in λ should be introduced explicitly when the decomposition (27) is presented in Section 3.","section":""}],"recommendation":"major_revision","confidential_remarks":"The paper's validation is entirely against the author's own earlier paper, and the two-level counterexample shows that the displayed Eq. (37) is wrong as printed. The error is a single exponent in the off-diagonal term and the trace-based derivation appears to support the corrected exponent, so the result is likely salvageable. I would urge the editor to require an independent numerical or exactly solvable check in the revision. The manuscript also leans heavily on self-citations; this is acceptable given the topic, but independent validation is especially important here."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Read this one with a red pen on eq. (37). The trace construction is a legitimate alternative route to rational-order sum rules, but the central displayed formula has a wrong exponent in the off-diagonal term. The derivation from eq. (34) to eq. (36) gives (epsilon_m^{1-s} - epsilon_n^{1-s})/(epsilon_m - epsilon_n); the printed (epsilon_m^{-s} - epsilon_n^{1-s})/(epsilon_m - epsilon_n) is not equivalent. I ran the two-level check: epsilon_1=1, epsilon_2=4, sigma=t(|1><2|+|2><1|), s=3/2. The exact second-order coefficient from the generalized eigenvalue problem is t^2/4. Eq. (37) as printed gives 11t^2/32; with the corrected exponent it gives t^2/4. So the paper's repeated claim that eq. (37) reproduces eq. (9) of Ref. [1] cannot be true as submitted.\n\nThat is the main finding. The rest of the paper is better than the typo suggests. The definition of fractional Green's functions of order 1/N and the iterative solution of the matrix equation are clean; routing Z(1/N+1/N') through traces is a real methodological step and avoids Rayleigh-Schrodinger degeneracy bookkeeping. The method extends to higher order, and the appendix gives those corrections. This is a re-derivation rather than a new sum rule -- the final expression duplicates Ref. [1] -- but an alternative derivation can still have value if it makes higher orders easier.\n\nSofter concerns: \"exact\" in the title overstates a second-order perturbative result; the perturbation series and the trace interchanges are not justified; validation is entirely against the author's own earlier formula; and there are smaller typos (eq. (3) should have sqrt(Sigma(y)), and there is a dangling eq. (??)). None of those is fatal once the exponent is fixed. The self-citation is not itself a flaw -- the problem is that this paper provides no independent check.\n\nRecommendation: this deserves a referee, not a desk reject. The referee should require the exponent correction, an explicit statement that the result is second-order in the density contrast, and ideally one independent check (even a two-level or finite-matrix example). After that, the paper is a serviceable contribution for people working on spectral sum rules for inhomogeneous systems. I would not cite it in its current form.","headline":"Eq. (37) has a wrong exponent in its off-diagonal term, so the paper's central claim to reproduce Ref. [1] fails as printed; the trace-based method is still coherent and worth a referee after a two-character fix.","tokens_in":9904,"tokens_out":14781,"would_cite":false,"duration_ms":129711,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35P20","35J05"],"pacs":[],"model":"deepseek-v4-flash","headline":"Rational-order spectral sum rules for Dirichlet billiards reduce to traces of products of fractional Green's functions","keywords":["spectral zeta function","sum rules","quantum billiards","Helmholtz equation","fractional Green's function","perturbation theory","trace identity","positive definite spectrum"],"falsifier":"For a one-dimensional inhomogeneous string with a known exact spectrum, such as a Dirichlet string with Σ(x)=1+λx, compute the exact eigenvalues numerically for several small λ, evaluate Z(3/2) and Z(4/3), subtract the λ=0 term, and compare the coefficient of λ² with eq. (37); any mismatch beyond numerical precision would show the trace derivation does not hold as stated.","tokens_in":8690,"feed_emoji":"🧮","tokens_out":7917,"duration_ms":73510,"temperature":0.7,"pith_summary":"Spectral sum rules are weighted sums of inverse eigenvalues, Z(s) = Σ_n $E_n^{{-s}}$; the paper treats rational exponents s. It claims that for Helmholtz operators with a positive density and a positive definite spectrum, such as Dirichlet billiards, these rational-order sum rules can be computed exactly to second order in the density contrast by expressing them as traces of products of Green's functions of fractional order. The resulting formula, eq. (37), coincides with the second-order formula obtained earlier by Rayleigh-Schrödinger perturbation theory, eq. (9) of Ref. [1], so the trace route provides an independent derivation of that result. A sympathetic reader would care because this turns a spectral sum that normally requires the full eigenvalue problem into a finite perturbative trace calculation over the homogeneous eigenbasis, and it extends the classical integer-order trace method of Ref. [2] to rational exponents.","feed_headline":"Fractional Green's functions unlock exact rational-order sum rules","feed_subtitle":"A trace identity reproduces the 2012 perturbative formula for spectral zeta functions Z(s) at rational s","key_machinery":"The carrier of the argument is the fractional Green's function of order 1/N, the integral kernel whose N-fold convolution gives G=√Σ G0 √Σ; its spectral coefficients $q^{{[1/N]}}$_{nm} satisfy a convolution matrix equation. For a weak inhomogeneity these coefficients are expanded as power series in λ, with the auxiliary objects $Δ^{{[1/N]}}$, $η^{{[1/N]}}$, and $ξ^{{[1/N]}}$ collecting the eigenvalue denominators that appear at each order. The sum rule Z(1/N+1/N') is obtained as the trace of a product of two such coefficient matrices, which turns the spectral sum into a trace that can be evaluated perturbatively.","core_discovery":"Writing the density as Σ(x)=1+λσ(x) with |σ|≪1, the paper defines G=√Σ G0 √Σ, where G0 is the homogeneous Green's function, and then introduces a fractional Green's function \\tilde $G^{{[1/N]}}$ whose N-fold convolution reproduces G. Expanding its spectral coefficients $q^{{[1/N]}}$_{nm} in powers of λ yields explicit second-order expressions built from the eigenvalue ratios encoded in Δ, η, and ξ. The sum rule at exponent s=1/N+1/N' is then evaluated as the trace of $q^{{[1/N]}}$ $q^{{[1/N']}}$, and the same universal expression is obtained for s=1+1/N. The final second-order sum rule is Z(s)=Σ_n $ε_n^{{-s}}$[1+λs⟨n|σ|n⟩ + (λ²/2)s(s-1)⟨n|σ|n⟩² + ...] - (λ²/2)s Σ_{n≠m}($ε_m^{{-s}}$-$ε_n^{{1-s}}$)/(ε_m-ε_n)⟨n|σ|m⟩⟨m|σ|n⟩ + O(λ³), which agrees with eq. (9) of Ref. [1]. Thus the paper establishes that the trace representation reproduces the earlier perturbative formula without invoking Rayleigh-Schrödinger perturbation theory, and it does so uniformly for all rational orders of the stated form.","pith_inferences":["The paper does not report a numerical test; one could directly check eq. (37) against exactly solvable one-dimensional strings, for instance Σ(x)=1+λx with Dirichlet conditions, by computing the coefficient of λ² in Z(3/2) numerically.","Because the same trace identity is algebraic once the fractional Green's function and the basis are known, the method should extend to other positive elliptic operators, such as weighted Laplacians on manifolds or graphs, provided a homogeneous Green's function is available.","The equality of the second-order expression for s=1+1/N and s=1/N+1/N' suggests that the perturbative sum rule depends only on s, not on the particular fractional decomposition, raising the question whether this s-only dependence holds at all orders and even for complex s in the convergence region.","The resummation leading to q≈Δ⟨n|√Σ|m⟩ hints that the diagonal part of Z(s) may be expressible as the homogeneous sum rule at shifted eigenvalues, but this is not proven in the paper."],"forward_implications":["Spectral zeta functions at rational s for Dirichlet Helmholtz problems can be approximated to second order in the density contrast using only eigenfunction matrix elements of the homogeneous problem, not the perturbed eigenfunctions or eigenvalues.","Because the method avoids eigenvalue perturbation series, the degeneracies that complicate Rayleigh-Schrödinger theory do not appear in this trace route.","In two dimensions, where Z(s) diverges for s ≤ 1, the rational-order expression permits s to approach 1 from above by taking N large, making the sum rules sensitive to the asymptotic spectrum and to corrections to Weyl's law.","The same calculation can in principle be pushed to higher order in λ; for order 1/2 the paper lists perturbative corrections to the fractional Green's function coefficients up to order eight.","For spectra containing a zero eigenvalue the trace representation is formally divergent, and the paper states that this case requires a separate renormalized treatment."],"supporting_citations":[{"why":"Supplies the reference result, eq. (9), a second-order perturbative formula for Z(s) that the trace calculation reproduces.","marker":"[1]"},{"why":"Introduces the trace-based sum-rule method for integer exponents in two-dimensional Dirichlet billiards, the technique this paper generalises to rational exponents.","marker":"[2]"},{"why":"Establishes the exact trace identity approach for inhomogeneous strings, the setting this paper extends.","marker":"[10]"},{"why":"Treats the null-eigenvalue case and shows the renormalization needed when traces are formally divergent, which the present paper explicitly excludes.","marker":"[12]"}],"fun_headline_variants":["Trace identity yields exact rational-order sum rules without perturbation","Fractional Green's functions reproduce spectral sum rules exactly","Perturbation-free rational sum rules via trace method","Uniform trace proof for rational-order quantum billiard sum rules"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The calculation assumes the density contrast is uniformly small, |σ(x)|≪1, and that truncating the perturbative expansion of the fractional Green's function at second order can be interchanged with the infinite trace sums; the paper does not prove convergence of either step.","fun_headline_variants_meta":{"raw":{"variants":["Trace identity yields exact rational-order sum rules without perturbation","Fractional Green's functions reproduce spectral sum rules exactly","Perturbation-free rational sum rules via trace method","Uniform trace proof for rational-order quantum billiard sum rules"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000621,"raw_usage":{"total_tokens":2859,"prompt_tokens":909,"completion_tokens":1950,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":525,"completion_tokens_details":{"reasoning_tokens":1885}},"tokens_in":525,"tokens_out":1950,"duration_ms":15928,"temperature":1.0,"reasoning_tokens":1885,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T11:36:27.520837+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"For a one-dimensional inhomogeneous string with a known exact spectrum, such as a Dirichlet string with Σ(x)=1+λx, compute the exact eigenvalues numerically for several small λ, evaluate Z(3/2) and Z(4/3), subtract the λ=0 term, and compare the coefficient of λ² with eq. (37); any mismatch beyond numerical precision would show the trace derivation does not hold as stated.","supporting_citations":[{"cited_title":"”A perturbative approach to the spectral zeta functions of strings, drums, and quantum billiards.” Journal of Mathematical Physics 53.12 (2012): 123519","cited_arxiv_id":null,"evidence_quote":"Supplies the reference result, eq. (9), a second-order perturbative formula for Z(s) that the trace calculation reproduces."},{"cited_title":"Moussa, and J","cited_arxiv_id":null,"evidence_quote":"Introduces the trace-based sum-rule method for integer exponents in two-dimensional Dirichlet billiards, the technique this paper generalises to rational exponents."},{"cited_title":"”Exact sum rules for inhomogeneous strings.” An - nals of Physics 338 (2013): 341-360","cited_arxiv_id":null,"evidence_quote":"Establishes the exact trace identity approach for inhomogeneous strings, the setting this paper extends."},{"cited_title":"”Exact sum rules for inhomogeneous systems co n- taining a zero mode.” Annals of Physics 349 (2014): 253-267","cited_arxiv_id":null,"evidence_quote":"Treats the null-eigenvalue case and shows the renormalization needed when traces are formally divergent, which the present paper explicitly excludes."}],"review_version":1}