{"id":"962698ae-8f2e-4541-87e8-e06f5978432d","arxiv_id":"1908.05392","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Explicit Krein resolvent identities for singular Sturm-Liouville operators are derived and applied to Bessel operators, yielding trace formulas and an explicit spectral shift function for ν in [0,1).","lead":"This paper derives explicit Krein resolvent identities for singular Sturm-Liouville operators, expressing the resolvent difference of two self-adjoint extensions as a rank-one or rank-two operator written in terms of boundary condition bases and the Lagrange bracket. It applies these identities to the Bessel operator on the half-line, producing explicit trace formulas and the spectral shift function for the Friedrichs extension paired with any other self-adjoint realization.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Most load-bearing input is the cited spectral assumption (5.6); internal identities and explicit Bessel calculations appear sound.","rationale":"The reader's conditional verdict identifies the cited spectrum of T_0^{(nu)} as the weakest assumption, and that is indeed the most load-bearing point: the SSF reconstruction and every negative-eigenvalue conclusion are anchored to the statement that the reference extension has spectrum [0,infinity) and no point spectrum. I independently re-checked the main internal steps and found no algebraic defect: the rank-one form in Theorem 3.4 follows from Lemma 3.2 with the correct sign, the Bessel inner products and trace formulas are consistent with the nu=1/2 special case, and the omitted nu=0 bracket calculation in (5.109) checks out from the standard asymptotics of J0 and Y0. The substantial algebra in Theorems 4.4--4.7, though deferred in places to inspection, is not needed for the Bessel application. Thus the only genuine risk is that the external spectral input (5.6) is mis-applied or mis-cited; if it is correct, the central claims of the paper hold. Since this is a standard cited result and the internal computations corroborate the final formulas, the reader's conditional verdict should remain unchanged rather than being upgraded or downgraded.","tokens_in":52057,"tokens_out":30054,"duration_ms":289933,"concrete_test":"For nu=0.3 and nu=0, test Eq. (5.6) directly: compute the Weyl--Titchmarsh m-function of the Friedrichs extension from the explicit solutions in (5.25) and (5.108), or equivalently shoot for L^2 solutions of tau_nu y=lambda y satisfying the theta=0 boundary condition at x=0 for real lambda<0. If any square-integrable solution is found, Eq. (5.6) fails and the spectral shift function normalization shifts; if none is found, the reference spectrum is as cited and the eigenvalue conclusions stand.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central construction is internally consistent: Theorem 3.4's rank-one resolvent identity (3.13) follows from the boundary-data identities in Lemma 3.2, and the trace formula (3.15) is correct. The mu=1/2 check reduces to the known Dirichlet--Neumann spectral shift function, and the nu=0 bracket computation omitted in (5.109), when performed directly from the J0/Y0 asymptotics, reproduces the stated constant ln(z^{1/2}/2)+gamma-i*pi/2. The single load-bearing step is external: Eq. (5.6), cited from [13], asserts that the reference extension T_0^{(nu)} is the Friedrichs extension with sigma=sigma_ac=sigma_ess=[0,infinity) and sigma_p=empty. The spectral shift function normalization (5.9), the reconstruction (5.50)--(5.51), and all eigenvalue conclusions via Lemma A.3 depend on this input. If the parametrization in [13] used a different boundary condition, or if T_0^{(nu)} had a hidden negative eigenvalue for some nu in [0,1), the baseline xi=0 near -infinity would shift and the eigenvalue claims (5.44)--(5.46) and (5.117) would be off. This is a standard theorem, so the risk is low, but it is exactly the assumption on which the Bessel applications rest.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper derives explicit Krein resolvent identities for singular Sturm–Liouville operators, distinguishing the cases of one and two limit-circle endpoints. For one limit-circle endpoint, the resolvent difference between a reference extension T_0 and any other extension T_θ is shown to be a rank-one operator built from the Weyl–Titchmarsh solution w_z, with an explicit trace formula (Theorem 3.4). For two limit-circle endpoints, analogous rank-two and rank-one identities are proved for separated and coupled boundary conditions (Theorems 4.4–4.7), and the regular-interval identities of [9] are recovered as special cases. The main application is the Bessel expression with ν∈[0,1): the authors compute the resolvent difference and its trace relative to the Friedrichs extension, then invert the trace formula to obtain explicit spectral shift functions and to locate the unique negative eigenvalue of non-nonnegative realizations (Propositions 5.1, 5.4, 5.8, 5.9).","tokens_in":52266,"tokens_out":41274,"duration_ms":384818,"significance":"If the formulas are taken with the intended bilinear pairing, the paper provides genuinely explicit, parameter-free Krein identities for singular endpoints and furnishes the first fully explicit spectral shift functions for the Bessel family in the regime ν∈[0,1), including the ν=0 case. The recovery of the regular theory from Section 4 and the reduction of the ν=1/2 case to the Dirichlet–Neumann spectral shift function are useful consistency checks. The algebraic structure of the proofs is detailed, and the trace formulas are concrete enough for direct verification; I independently spot-checked the ν=1/2 and ν=0 computations. The main reservation is that the notation for the pairing in the rank-one terms is inconsistent with the declared Hilbert-space inner product, which affects the literal statement of the principal theorems; this is fixable by a notational revision, not by new mathematics.","major_comments":[{"comment":"The symbol ⟨w_z,·⟩_(a,b) is used both as the Hilbert-space inner product declared in §1 (linear in the second argument, hence conjugate-linear in the first) and as the bilinear pairing f↦∫_a^b w_z f r dx. Under the declared inner product, the rank-one operator in (3.13) has kernel w_z(x) overline{w_z(y)}, and its Hilbert-space trace would be k_θ(z)^{-1}‖w_z‖², which is real and nonnegative. But the Bessel computations in (5.29) and (5.111), and the resulting trace formulas (5.35) and (5.116), use the bilinear expression ∫ w_z², which for non-real z is complex; for instance the ν=1/2, θ=π/2 case gives trace -1/(2z), not a positive real number. The proof of Lemma 3.2 and the derivation of (3.13) also rely on the boundary value ∫ w_z f, not ∫ overline{w_z} f. Thus, as written, the principal resolvent identities are not literally correct with the stated inner product, although they are correct if ⟨·,·⟩_(a,b) is read as the bilinear form (f,g)↦∫ f g r dx. Please introduce a distinct notation for this bilinear pairing and restate Theorems 3.4, 4.4–4.7, and the trace formulas using it; alternatively, reformulate the rank-one terms using the Hilbert inner product with w_\\bar z in the first slot, where w_\\bar z=overline{w_z}.","section":"§3, Eq. (3.13); also (3.15), (4.21), (4.52), (4.69), (4.108)"},{"comment":"The identity [w_{z,0},ψ_{0,0}](0)=ln(z^{1/2}/2)+γ−iπ/2 is stated with the parenthetical remark “a calculation, omitted here.” This bracket determines k_{θ,0}(z), the trace formula (5.116), and ultimately the spectral shift function (5.117) and the negative eigenvalue claim for the ν=0 family. Because this is a load-bearing step, the computation should be included in the paper or an explicit reference supplied. I verified independently from the J0/Y0 small-argument asymptotics that the stated value is correct, so this is a local gap rather than an error.","section":"§5.2, Eq. (5.109)"}],"minor_comments":[{"comment":"The Bessel applications depend on the cited result from [13] that T_0^{(ν)} is the Friedrichs extension with purely absolutely continuous spectrum [0,∞) and empty point spectrum. Please state explicitly which boundary condition in the parametrization of [13] corresponds to θ=0 and confirm that no hidden negative eigenvalue occurs for any ν∈[0,1), so that the SSF normalization (5.9) and the eigenvalue conclusions via Lemma A.3 are unambiguous.","section":"§5, Eq. (5.6)"},{"comment":"The assertion that T_θ^{(ν)} and T_0^{(ν)} are bounded from below is used to invoke the standard existence and uniqueness of the spectral shift function, but no proof or reference is given at that point. Since this is a standard consequence of finite deficiency indices and semiboundedness of the Friedrichs extension, a one-sentence citation would suffice.","section":"§5.1, after Eq. (5.7)"},{"comment":"There are numerous typographical and formatting artifacts in the LaTeX source, such as “/greaterorequalslant”, “0<ε≪1” being typeset inconsistently, and missing overlines around conjugate quantities. These should be cleaned up in the final version.","section":"Throughout"},{"comment":"If the recommended change to a bilinear pairing is adopted, Lemma 3.3 must be supplemented with the corresponding statement for operators of the form A=(φ,·)ψ, where (φ,g)=∫ φ g r dx; the trace is then (φ,ψ)=∫ φ ψ r dx, not ⟨φ,ψ⟩_H.","section":"Lemma 3.3"},{"comment":"The recovery of the regular-interval identities from [9] is convincing but somewhat compressed; explicitly displaying the index interchange in (4.138) would make the comparison easier to verify.","section":"Remark 4.9"}],"recommendation":"major_revision","confidential_remarks":"The mathematical content appears sound under the intended convention, and the Bessel spectral shift functions are valuable explicit results. The blocking issue is the notational collision between the Hilbert inner product and the bilinear pairing in the rank-one resolvent identities; this affects the literal statement of the paper's main theorems and must be repaired before publication. The omitted calculation in (5.109) and the reliance on (5.6) are secondary and easily addressed."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nMy take: this is a solid, useful piece of spectral theory, and the Bessel application is the part that will get cited. They take Krein's resolvent identity, previously written out for regular endpoints, and make it explicit for singular Sturm-Liouville operators with one or two limit circle endpoints, using boundary condition bases and Lagrange brackets. The identities in Theorems 3.4 and 4.4–4.7 look right, and the rank-one/trace-class structure falls out naturally. I spot-checked the ν=1/2 reduction and the omitted ν=0 bracket in (5.109); both come out as stated. That gives me real confidence in the algebra.\n\nWhat is new here is not the abstract Krein formalism—that is a known shape—but the explicit singular-endpoint form and the concrete Bessel payoff: trace formulas (5.35) and (5.116) and the spectral shift functions (5.42)–(5.46) and (5.117). The SSF is obtained by inverting a trace formula, not by fitting, and it recovers the known characterizations of nonnegative Bessel realizations (Corollaries 5.5 and 5.10) plus the single negative eigenvalue. That is a genuinely reusable result.\n\nSoft spots are minor but real. The whole Bessel section leans on the cited result (5.6) from Everitt–Kalf that T_0^(ν) is the Friedrichs extension with purely absolutely continuous spectrum [0,∞) and no eigenvalues. If that input were wrong, the SSF normalization and eigenvalue conclusions would shift. I think the risk is low—this is a standard theorem—but it is load-bearing, and a referee should ask the authors to state it as a hypothesis or verify it explicitly. The Green's function representations are also imported from [11], and two places in Section 4 wave at algebra 'by inspection.' None of this undermines the main argument, but those are the spots to fill in.\n\nCitation overlap with the authors' earlier work is real but appropriate; they are extending [9] and [11], not relabeling them.\n\nWho is this for? Spectral theorists working with Krein resolvent formulas, singular Sturm-Liouville operators, or Bessel operators will get direct use out of it. I'd send it to a serious referee and expect acceptance after modest revision.","headline":"Solid, reusable paper: explicit Krein identities for singular Sturm-Liouville endpoints plus Bessel trace formulas and spectral shift functions; the Bessel application hinges on one standard cited spectral fact.","tokens_in":52849,"tokens_out":4515,"would_cite":true,"duration_ms":42074,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["47A10","47A55","47A56","47B10"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper proves the resolvent difference for one-limit-circle Sturm-Liouville operators is a rank-one kernel $k_\\theta(z)^{-1}\\langle w_z,\\cdot\\rangle w_z$, and derives explicit Bessel spectral shift functions and the lone negative…","keywords":["Krein resolvent identity","singular Sturm-Liouville operators","Bessel operators","spectral shift function","Weyl-Titchmarsh solution","Lagrange bracket","boundary condition bases","rank-one resolvent difference"],"falsifier":"For fixed $\\nu\\in(0,1)$ and $\\theta\\in(\\pi/2,\\pi)$, solve the Bessel eigenvalue equation at negative energy with the boundary condition $\\cos(\\theta)[y,\\varphi_{0,\\nu}](0)+\\sin(\\theta)[y,\\psi_{0,\\nu}](0)=0$ and compare the computed eigenvalue with $e_{\\theta,\\nu}$ from (5.39); a mismatch would disprove the spectral shift reconstruction. For $\\nu=0$, the analogous comparison uses $e_{\\theta,0}=-4e^{-2[\\cot(\\theta)+\\gamma]}$ from (5.117).","tokens_in":51825,"feed_emoji":"📉","tokens_out":10753,"duration_ms":90723,"temperature":0.7,"pith_summary":"The paper derives explicit Krein resolvent identities for singular Sturm-Liouville operators when one endpoint is limit circle and the other is limit point, and also for two limit circle endpoints. In the one-limit-circle case it proves that the resolvent difference of any two self-adjoint extensions is a rank-one operator of the form $k_\\theta(z)^{-1}\\langle w_z,\\cdot\\rangle w_z$, where $w_z$ is the Weyl-Titchmarsh solution, so its trace is $\\langle w_z,w_z\\rangle/k_\\theta(z)$. Applying this to the Bessel expression $-d^2/dx^2+(\\nu^2-1/4)x^{-2}$ on $(0,\\infty)$ for $\\nu\\in[0,1)$, it computes the trace of the resolvent difference explicitly and uses it to determine the spectral shift function for the Friedrichs extension versus any other self-adjoint realization. The resulting spectral shift functions locate the single simple negative eigenvalue that appears exactly when the realization is not nonnegative, and they recover the known characterization of nonnegative Bessel realizations.","feed_headline":"Bessel resolvent differences are rank-one and explicit","feed_subtitle":"Exact spectral shift functions follow; every non-nonnegative realization has one negative eigenvalue.","key_machinery":"The load-bearing objects are boundary condition bases $\\{\\varphi,\\psi\\}$ at a limit-circle endpoint, normalized by $[\\psi,\\varphi](c)=1$, and the Lagrange bracket $[f,g](x)=f(x)(pg')(x)-(pf')(x)g(x)$, which provides finite boundary data at singular endpoints where ordinary boundary values may not exist. The argument also uses the Weyl-Titchmarsh solution $w_z$, the unique solution of $\\tau y=zy$ square-integrable near the limit-point endpoint and normalized by $[w_z,\\varphi_a](a)=1$. The scalar $k_\\theta(z)=\\cot(\\theta)+[w_z,\\psi_a](a)$ measures how the boundary condition at $a$ changes, and the rank-one kernel $\\langle w_z,\\cdot\\rangle w_z$ carries the whole resolvent difference. In the two-limit-circle case the same machinery produces a $2\\times2$ matrix $K_{\\alpha,\\beta}(z)$ or $K_{R,\\eta}(z)$ and a rank-two (sometimes rank-one) correction.","core_discovery":"The central claim is Theorem 3.4: for a singular Sturm-Liouville expression with exactly one limit circle endpoint $a$, a fixed reference extension $T_0$, and another extension $T_\\theta$ parametrized by a boundary condition basis $\\{\\varphi_a,\\psi_a\\}$, for $z\\in\\rho(T_0)\\cap\\rho(T_\\theta)$, $(T_\\theta-zI)^{-1}-(T_0-zI)^{-1}=k_\\theta(z)^{-1}\\langle w_z,\\cdot\\rangle w_z$, with $k_\\theta(z)=\\cot(\\theta)+[w_z,\\psi_a](a)$, where $w_z$ is the Weyl-Titchmarsh solution. Hence the resolvent difference is rank one and its trace is $\\langle w_z,w_z\\rangle/k_\\theta(z)$. For the Bessel expression the paper evaluates all ingredients in terms of Bessel and Hankel functions, obtaining the trace formulas (5.35) for $\\nu\\in(0,1)$ and (5.116) for $\\nu=0$, and then inverts the Stieltjes transform to obtain the spectral shift functions (5.42)-(5.46) and (5.117). These formulas show which realizations are nonnegative and expose the single negative eigenvalue $e_{\\theta,\\nu}$ for non-nonnegative realizations.","pith_inferences":["The same rank-one identity should apply to other singular one-limit-circle Sturm-Liouville expressions, such as radial Schrödinger operators with Coulomb-like singular terms, wherever the Weyl-Titchmarsh solution is known in closed form.","Because the resolvent difference is literally a rank-one term, a cheap numerical test of the spectral shift reconstruction is to evaluate $\\langle w_z,w_z\\rangle/k_\\theta(z)$ at a few complex energies and compare the predicted negative eigenvalue with a direct shooting method.","The logarithmic term in $k_{\\theta,0}(z)$ for $\\nu=0$ suggests that the critical case may share features with other operators having borderline long-range potentials, beyond the Bessel example treated here."],"forward_implications":["For any singular Sturm-Liouville pair with one limit-circle endpoint, the full resolvent difference is a single explicit rank-one operator, so trace identities and spectral shift functions follow without separately constructing Green's functions.","For Bessel operators with $\\nu\\in(0,1)$, the Friedrichs extension and any other realization are resolvent comparable, and their spectral shift function is given by the arctangent formulas (5.42)-(5.46), which simultaneously characterize $\\theta\\in[0,\\pi/2]$ as exactly the nonnegative realizations.","For $\\nu=0$, the unique nonnegative realization is $\\theta=0$, and every other realization has exactly one simple negative eigenvalue $e_{\\theta,0}$; in the special case $\\nu=1/2$, the trace identity for the Neumann minus Dirichlet Laplacian on the half-line follows.","The explicit negative eigenvalues $e_{\\theta,\\nu}$ and $e_{\\theta,0}$ are determined directly from $\\theta$, $\\nu$, and the gamma function, so the spectrum of any real Bessel realization is exactly $[0,\\infty)$ plus at most one negative eigenvalue."],"supporting_citations":[{"why":"Derives Krein resolvent identities for regular Sturm-Liouville operators; the singular identities in this paper are designed to recover and generalize those formulas as special cases.","marker":"[9]"},{"why":"Supplies the Weyl-Titchmarsh theory, Lagrange bracket identities, endpoint classification, and Green's function representation used in the proof of Theorem 3.4.","marker":"[11]"},{"why":"Provides the Bessel boundary condition bases and the spectral fact that $T_0^{(\\nu)}$ has purely absolutely continuous spectrum $[0,\\infty)$ with no eigenvalues, the reference-spectrum assumption for the application.","marker":"[13]"},{"why":"Establishes which Bessel realizations are Friedrichs or Krein-von Neumann extensions and gives the characterization of nonnegative extensions that the spectral shift results reproduce.","marker":"[4]"},{"why":"Gives the definition of boundary condition bases and the parametrization of self-adjoint extensions required for the resolvent identities.","marker":"[26]"},{"why":"Gives the abstract spectral shift function existence and trace formula used to invert the resolvent trace.","marker":"[22]"},{"why":"Provides the Stieltjes inversion technique used to recover the spectral shift function from boundary values of the logarithmic trace.","marker":"[17]"}],"fun_headline_variants":["Rank-one resolvent differences yield Bessel spectral shifts","Explicit trace and spectral shift for Bessel operators","Bessel resolvent rank-one: exact spectral shifts","Krein resolvent identities for Bessel: rank-one traces"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The reconstruction assumes the reference extension $T_0^{(\\nu)}$ has spectrum exactly $[0,\\infty)$ with no eigenvalues, and that both operators are bounded below with trace-class resolvent difference; if hidden negative spectrum or an embedded eigenvalue existed, the eigenvalue conclusions would change.","fun_headline_variants_meta":{"raw":{"variants":["Rank-one resolvent differences yield Bessel spectral shifts","Explicit trace and spectral shift for Bessel operators","Bessel resolvent rank-one: exact spectral shifts","Krein resolvent identities for Bessel: rank-one traces"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000226,"raw_usage":{"total_tokens":1454,"prompt_tokens":915,"completion_tokens":539,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":531,"completion_tokens_details":{"reasoning_tokens":474}},"tokens_in":531,"tokens_out":539,"duration_ms":5468,"temperature":1.0,"reasoning_tokens":474,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:15:30.168920+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"For fixed $\\nu\\in(0,1)$ and $\\theta\\in(\\pi/2,\\pi)$, solve the Bessel eigenvalue equation at negative energy with the boundary condition $\\cos(\\theta)[y,\\varphi_{0,\\nu}](0)+\\sin(\\theta)[y,\\psi_{0,\\nu}](0)=0$ and compare the computed eigenvalue with $e_{\\theta,\\nu}$ from (5.39); a mismatch would disprove the spectral shift reconstruction. For $\\nu=0$, the analogous comparison uses $e_{\\theta,0}=-4e^{-2[\\cot(\\theta)+\\gamma]}$ from (5.117).","supporting_citations":[{"cited_title":"Clark, F","cited_arxiv_id":null,"evidence_quote":"Derives Krein resolvent identities for regular Sturm-Liouville operators; the singular identities in this paper are designed to recover and generalize those formulas as special cases."},{"cited_title":"Eckhardt, F","cited_arxiv_id":null,"evidence_quote":"Supplies the Weyl-Titchmarsh theory, Lagrange bracket identities, endpoint classification, and Green's function representation used in the proof of Theorem 3.4."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the Bessel boundary condition bases and the spectral fact that $T_0^{(\\nu)}$ has purely absolutely continuous spectrum $[0,\\infty)$ with no eigenvalues, the reference-spectrum assumption for the application."},{"cited_title":"Ananieva and V","cited_arxiv_id":null,"evidence_quote":"Establishes which Bessel realizations are Friedrichs or Krein-von Neumann extensions and gives the characterization of nonnegative extensions that the spectral shift results reproduce."},{"cited_title":"Zettl, Sturm–Liouville Theory , Mathematical Surveys and Monographs, Vol","cited_arxiv_id":null,"evidence_quote":"Gives the definition of boundary condition bases and the parametrization of self-adjoint extensions required for the resolvent identities."},{"cited_title":"Schm¨ udgen,Unbounded Self-Adjoint Operators on Hilbert Space, Graduate Texts in Math- ematics, Springer, New York, 2012","cited_arxiv_id":null,"evidence_quote":"Gives the abstract spectral shift function existence and trace formula used to invert the resolvent trace."},{"cited_title":"Gesztesy and E","cited_arxiv_id":null,"evidence_quote":"Provides the Stieltjes inversion technique used to recover the spectral shift function from boundary values of the logarithmic trace."}],"review_version":1}