{"id":"40fb4be7-0940-4d46-8a40-6c3db728c83c","arxiv_id":"2605.31368","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For semibounded self-adjoint operators with compact resolvent, a generic bounded or smooth potential breaks all eigenvalue multiplicities whenever a finite-dimensional splitting condition holds; applications include sub-Laplacians and rough-domain Laplacians.","lead":"An abstract theorem gives conditions under which adding a generic bounded potential to a self-adjoint operator makes every eigenvalue simple. The authors apply it to sub-Laplacians, maximally hypoelliptic operators, and Laplacians on rough domains, yielding generic simplicity and Courant-type nodal bounds.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Bounded-domain magnetic examples state Q=C∞(Ω,R), which is not continuously embedded in L∞; the abstract theorem's admissibility premise fails for these applications as written.","rationale":"The reader's verdict is CONDITIONAL because the abstract theorem appears correct while the bounded-domain magnetic Schrödinger applications use a perturbation space that does not satisfy the theorem's admissibility hypothesis. I agree. I checked the main proof: Lemma 2.7 gives L∞-Lipschitz continuity of eigenvalues via min–max; Lemma 2.9 gives openness of A_n; Lemma 2.8 gives real-analytic eigenvalue branches for the family L_p+tM_σ by Kato's theory; Lemma 2.10 uses only the non-scalarity of G_σ to split a multiple eigenvalue, and the multiplicity drops by at least one, so Lemma 2.11's finite iteration is valid. Proposition 2.3, which supplies the non-scalar perturbation from C_c^∞, is also correct. Thus Theorem 2.13 itself is not endangered. The remaining issue is genuinely load-bearing for the paper's advertised scope: several applications, especially Examples 3.11 and 3.12, state the perturbation space as C∞(Ω,R) on a bounded open set. Such functions need not be bounded, so M_q is not a bounded perturbation and Q is not admissible under Definition 2.1. This is a fixable but real gap in the applications, not an internal inconsistency of the central argument. A concrete check using q(x)=tan(π(x−1/2)) on (0,1) settles the failure of the embedding; rerunning the example with C_b^∞(Ω,R) or L∞(Ω,R) would restore applicability. The reader already identified this as the weakest assumption, so I mark agreement as 'agree.' The appropriate disposition is UNCHANGED: the CONDITIONAL verdict remains.","tokens_in":23510,"tokens_out":15423,"duration_ms":158577,"concrete_test":"Take Ω=(0,1) and q(x)=tan(π(x−1/2)). Then q∈C∞(Ω,R) but sup|q|=∞, so the continuous embedding C∞(Ω,R)↪L∞(Ω,R) required by Definition 2.1 fails. Then check whether Examples 3.11/3.12 can be rerun verbatim with Q replaced by C_b^∞(Ω,R) (or C∞(Ω̄,R) for regular Ω) and whether the cited compact-resolvent and self-adjointness results from [26] remain valid for all bounded potentials in that space. If yes, the only correction is the stated perturbation space, confirming the reader's CONDITIONAL verdict.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The core Baire-category argument (Lemmas 2.9–2.11 and Theorem 2.13) is internally sound; I find no flaw in the finite-dimensional splitting argument or in the analytic perturbation step. The load-bearing premise is admissibility in Definition 2.1: Q must be a Fréchet space continuously embedded in L∞(X,R), and Hypothesis (H) must be verified. Proposition 2.3 reduces the latter to C_c^∞(X°,R) ⊂ Q. This is delicate in the bounded-domain applications. In Examples 3.11 and 3.12 the perturbation space is stated as C∞(Ω,R) for a bounded open set Ω. For such Ω, C∞(Ω,R) contains unbounded functions (e.g. q(x)=1/dist(x,∂Ω)), so it is not contained in L∞(Ω,R), and the compact-open Fréchet topology does not make the inclusion into L∞ continuous. Thus the pair (L,Q) is not admissible and Theorem 2.13 does not apply verbatim. The abstract theorem survives; the applications need Q=C_b^∞(Ω,R), Q=C∞(Ω̄,R) for regular domains, or Q=L∞(Ω,R). This is exactly the reader's weakest assumption.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proves an abstract genericity theorem (Theorem 2.13) for simplicity of all eigenvalues of a self-adjoint, semibounded operator L with compact resolvent, under real bounded perturbations from an admissible Fréchet space Q. The proof combines a min–max Lipschitz estimate (Lemma 2.7), Kato-type analytic perturbation theory (Lemma 2.8), a finite-dimensional splitting lemma (Proposition 2.3), and a Baire-category argument. The perturbation space need not contain smooth eigenfunctions; the splitting is performed directly in L2. The abstract theorem is then applied to sub-Laplacians, maximally hypoelliptic operators, sums of squares with continuous coefficients, Dirichlet/Neumann/Robin Laplacians on bounded domains, magnetic Schrödinger operators, non-compact oscillators, and a Coulomb-type example with non-compact resolvent.","tokens_in":23753,"tokens_out":15803,"duration_ms":133894,"significance":"If correct, the abstract theorem provides a unified explanation of residual spectral simplicity for a much broader class than the classical elliptic setting. The proof avoids unique continuation entirely for the global result, which is a notable simplification. The applications are numerous and connect to current work on sub-Riemannian nodal theorems. The paper is mostly well organized and the central argument is transparent. The main weakness is a mismatch between the abstract admissibility hypothesis and the perturbation spaces stated in two bounded-domain magnetic examples, plus a few smaller presentation gaps. These are fixable and do not undermine the core theorem.","major_comments":[{"comment":"The perturbation space is stated as Q = C^∞(Ω,R) for a bounded set Ω. If Ω is an open bounded set, C^∞(Ω,R) contains unbounded functions and is not contained in L^∞(Ω,R); the compact-open Fréchet topology does not give a continuous embedding into L^∞. Thus the pair (L,Q) is not admissible in the sense of Definition 2.1 and Theorem 2.13 does not apply as written. The two magnetic Schrödinger applications need Q = C_b^∞(Ω,R), or Q = C^∞(Ω̄,R) with Ω a regular bounded domain and this convention stated explicitly. This is a load-bearing issue for these examples.","section":"Examples 3.11, 3.12 (§3.2.2)"},{"comment":"The abstract theorem is stated for L^2(X,dµ) with a fixed smooth density, but Proposition 3.6 applies it to the weighted space L^2_ρ(Ω) with only ρ ∈ L^∞, ρ ≥ ρ0 > 0. The sentence 'it can also be applied directly in the weighted space L^2_ρ(Ω)' is not a proof. The extension is straightforward because the argument uses only the Hilbert-space structure and C_c^∞(Ω) ⊂ Q, but it should be stated and justified explicitly, or Theorem 2.13 should be generalized to admissible positive densities.","section":"Proposition 3.6 (§3.2.1)"}],"minor_comments":[{"comment":"The statement has '∀c∈R' in the wrong place; it should read 'G_σ ≠ c I_E for every c∈R'. In the proof, the inner product ⟨M_σφ,ψ⟩ is written as ∫ σ φψ dμ; a complex conjugate on ψ should appear: ∫ σ φ ψ̄ dμ.","section":"Proposition 2.3"},{"comment":"The sentence 'By Lemma 2.9, each A_n is open in Q.' appears twice in succession; remove the duplicate.","section":"Theorem 2.13 proof"},{"comment":"The residual set is stated as a subset of C_b^∞(R^2,R), but the operator is on R^n. The dimension should be R^n (or the notation made consistent).","section":"Example 3.14"},{"comment":"The shorthand 'L+q' is used for L+M_q; for precision, especially when discussing the abstract theorem, use L_q or L+M_q consistently.","section":"Section 1.3 and throughout"},{"comment":"The electric potential V appears both as part of the base operator and as the perturbation parameter. It would help to state explicitly that the base operator is taken with V=0 and that V is the perturbation.","section":"Examples 3.11, 3.12"}],"recommendation":"major_revision","confidential_remarks":"The core abstract theorem is sound and the paper is worth publishing after revision. The main issue is the admissibility of C^∞(Ω,R) in the two bounded-domain magnetic examples; this is easily fixed but needs to be addressed before the applications can be accepted as stated. Also, the weighted L^2_ρ application in Proposition 3.6 should be formalized. I do not see any problem with the Baire-category argument or the analytic perturbation step."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nYou should know: the core of this paper is correct. The abstract Theorem 2.13—residual simplicity under a finite-dimensional splitting hypothesis—is proved cleanly via Kato analytic perturbation theory and Baire category, and the proof has no gaps I can find. The genuinely new piece is Proposition 2.3: if compactly supported smooth functions are available as perturbations, no unique continuation is needed to split any multiple eigenspace. That is what opens the door to sub-Laplacians, maximally hypoelliptic operators, and L∞ potentials on rough domains.\n\nThe applications are mostly credible. The sub-Laplacian and Courant nodal discussion, the maximally hypoelliptic case via [6], and the low-regularity Dirichlet/Neumann/Robin examples all check out as legitimate domains of application. The paper also gives credit where due: the splitting idea is Albert's, and the packaging is standard; the new content is the abstract criterion and its range.\n\nThe soft spots are real but minor. The stress-test note is accurate: in Examples 3.11 and 3.12, they take Q=C∞(Ω,R) for a bounded domain. On an open bounded set, C∞(Ω) is not contained in L∞, and the compact-open Fréchet topology does not make the inclusion continuous. The abstract theorem requires an admissible perturbation space. The fix is simple—take Q=C_b^∞(Ω) or C∞(Ω̄) for a regular domain, or Q=L∞(Ω)—but as written those two applications do not follow verbatim. The reader's \"weakest assumption\" is correctly identified.\n\nAlso: there are minor typos—a missing conjugate in the inner product in Proposition 2.3 (the argument still works with real test functions), and a duplicated sentence in the proof of Theorem 2.13. The dependence on deep external results ([6], [17], [11]) is worth noting but not a flaw; the core theorem is self-contained, and the applications explicitly rely on the cited theorems.\n\nFor a referee: I would send this out. The abstract theorem deserves publication, and the applications are valuable once the perturbation-space issue in the magnetic examples is fixed. The paper is honest, the reasoning is clear, and the central claim holds up.\n\nBest,\n[You]","headline":"Core abstract theorem is sound and the no-unique-continuation splitting criterion is a real advance; publication-worthy after fixing the C∞(Ω) vs. C_b^∞(Ω)/C∞(Ω̄) admissibility slip in the magnetic examples and a few typos.","tokens_in":24270,"tokens_out":3869,"would_cite":true,"duration_ms":34269,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["47A10","47A55","35P05","47B25"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper proves a general abstract criterion: for a semibounded self-adjoint operator with compact resolvent, a residual set of admissible bounded potentials makes every eigenvalue simple.","keywords":["generic simplicity","self-adjoint operators","bounded potentials","residual set","compact resolvent","sub-Laplacian","nodal-domain theorem","spectral perturbation"],"falsifier":"Find a multiple eigenspace E of some L_q and an admissible potential space Q satisfying Hypothesis (H) such that for all σ in a neighbourhood of zero in Q, the compression (P_E M_σ)|_E remains scalar. Proposition 2.3 proves this is impossible when C_c^∞⊂Q, so any concrete example—e.g., a double eigenvalue of the Dirichlet Laplacian on a rectangle with a small supported bump potential—would pinpoint the failure if the splitting were absent.","tokens_in":23354,"feed_emoji":"🔢","tokens_out":9042,"duration_ms":88544,"temperature":0.7,"pith_summary":"This paper proves a general criterion for generic simplicity of the spectrum of a self-adjoint operator under bounded potential perturbations. If the operator is semibounded with compact resolvent and the allowed potentials form a suitable complete metrisable space of bounded functions, then the set of potentials for which every eigenvalue is simple is residual in that space—meaning it contains a countable intersection of open dense sets and is itself dense. The proof isolates the mechanism from operator-specific tools such as ellipticity and unique continuation, so the criterion can be applied to sub-Laplacians, maximally hypoelliptic operators, Laplacians on rough bounded domains with several boundary conditions, and Schrödinger-type operators on non-compact spaces. A corollary is a generic nodal-domain theorem: for a residual set of potentials, every eigenfunction of the perturbed operator satisfies the classical nodal bound.","feed_headline":"Generic bounded potentials make all eigenvalues simple","feed_subtitle":"An abstract criterion covers sub-Laplacians, rough domains, and non-compact Schrödinger operators.","key_machinery":"The key object is Hypothesis (H), a finite-dimensional splitting condition: for each multiple eigenspace E of any L_q, there must exist an admissible potential σ whose projection onto E is non-scalar. This is the only operator-specific input needed; once it holds, analytic perturbation theory supplies real-analytic eigenvalue branches whose first derivatives are the eigenvalues of this compression, so a small perturbation in the non-scalar direction splits the eigenvalue. Proposition 2.3 shows that compactly supported smooth functions always provide such a σ, using a measure-theoretic argument, so C_c^∞(X∘,R)⊂Q suffices. The openness part uses the min–max principle and the continuous embeddi","core_discovery":"The central claim is Theorem 2.13: if L is self-adjoint, semibounded, with compact resolvent on L2(X), and the pair (L,Q) satisfies Hypothesis (H), then the set G_Q of potentials q∈Q for which every eigenvalue of L+M_q is simple is residual in Q. Hypothesis (H) is a finite-dimensional splitting condition: for every q and every eigenvalue λ of L_q whose eigenspace E has dimension at least two, there is a potential σ∈Q such that the compression of multiplication by σ to E is not a scalar multiple of the identity on E. The proof is a classical category argument: each set A_n of potentials that make the first n eigenvalues simple is open, by Lipschitz dependence of eigenvalues on the L∞ norm, an","pith_inferences":["Inference: Because the proof uses no unique continuation, generic simplicity should persist for sums-of-squares operators where localised splitting may fail; only the global version with all compactly supported potentials is needed.","Inference: The criterion suggests a recipe for proving generic simplicity in other perturbation spaces (e.g., band-limited or sparse potentials): check the finite-dimensional splitting condition and continuity of the embedding into L∞; the category argument then applies unchanged.","Inference: The localised version ties splitting to weak unique continuation; one could test whether generic simplicity holds for perturbations supported in an arbitrary open set ω in elliptic cases, where classical unique-continuation results apply."],"forward_implications":["For sub-Laplacians on compact manifolds, a residual set of smooth potentials makes every eigenvalue simple and, consequently, the nodal-domain theorem holds for all eigenfunctions of the perturbed operator.","For maximally hypoelliptic operators satisfying the algebraic characterisation of maximal hypoellipticity, residual simplicity follows whenever the operator is formally self-adjoint and bounded below.","For Dirichlet, Neumann, and Robin Laplacians on bounded domains with rough boundaries, the residual set of potentials lives in L∞(Ω), so no regularity up to the boundary is required.","For non-compact examples—harmonic and anharmonic oscillators, magnetic Schrödinger operators with compact resolvent—bounded smooth potentials generically give simple spectra.","If the operator only has discrete spectrum in a spectral region below the essential spectrum, the same category argument applies to that region; for example, the negative eigenvalues of the Coulomb Hamiltonian are generically simple."],"fun_headline_variants":["Generic bounded potentials force simple spectra","Bounded perturbations generically split eigenvalues","Simplicity of spectra for generic bounded potentials","All eigenvalues simple for typical bounded potentials","Generic potentials make every eigenvalue distinct"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The load-bearing premise is that the allowed potentials form a complete metrisable space continuously embedded in L∞ and rich enough to split every multiple eigenspace (e.g., containing all compactly supported smooth functions); if the embedding is not continuous, the openness of the simplicity sets is lost, and if the space is too small, the density step fails.","fun_headline_variants_meta":{"raw":{"variants":["Generic bounded potentials force simple spectra","Bounded perturbations generically split eigenvalues","Simplicity of spectra for generic bounded potentials","All eigenvalues simple for typical bounded potentials","Generic potentials make every eigenvalue distinct"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000161,"raw_usage":{"total_tokens":1035,"prompt_tokens":668,"completion_tokens":367,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":412,"completion_tokens_details":{"reasoning_tokens":307}},"tokens_in":412,"tokens_out":367,"duration_ms":4651,"temperature":1.0,"reasoning_tokens":307,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-02T12:44:55.303373+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Find a multiple eigenspace E of some L_q and an admissible potential space Q satisfying Hypothesis (H) such that for all σ in a neighbourhood of zero in Q, the compression (P_E M_σ)|_E remains scalar. Proposition 2.3 proves this is impossible when C_c^∞⊂Q, so any concrete example—e.g., a double eigenvalue of the Dirichlet Laplacian on a rectangle with a small supported bump potential—would pinpoint the failure if the splitting were absent.","supporting_citations":[],"review_version":2}