{"id":"b6e0ee1e-6237-4bd1-bc8f-56baaae6b575","arxiv_id":"2411.16494","paper_version":2,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"A new relativistic Dirac oscillator with a complex rotation has real discrete spectrum but wild eigenfunctions and pseudospectra; the eigenprojector growth rate equals a known rotated-oscillator quantity.","lead":"This paper introduces a relativistic (Dirac) version of the rotated harmonic oscillator, a famous non-self-adjoint operator used to model resonances. It proves the eigenvalues stay real and discrete, while the eigenfunctions misbehave and the pseudospectrum is highly non-trivial far from the spectrum.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Proposition 5.2's key estimate ||Bψ|| ≤ |tan(θ/2)| ||(A−i Im z)ψ|| is false for small |Im z|; the proof of Theorem 1.2's upper inclusion is therefore invalid, though Theorem 1.1 is unaffected.","rationale":"The paper's headline result, Theorem 1.1, is carefully derived from the known estimate (2.8), and my check of Lemmas 4.3–4.5 found no flaw: the upper bound via Schwarz and the lower bound via the eigenvector χ̃¹_n both work, and the asymptotic rate is m-independent. The lower pseudospectral inclusion in Theorem 1.2 also follows soundly from (2.9) through the block decomposition Hθ² = T+⊕T−. The problem is concentrated in the proof of the upper inclusion, which depends entirely on Proposition 5.2. The exact identities (3.2)–(3.3) do not justify the relative-bound inequality used there; a direct computation with the zero mode of A gives a counterexample to the displayed estimate while Proposition 5.2's hypothesis is still satisfied. This is a genuine, localized gap: it does not affect Theorem 1.1 or the non-triviality of the pseudospectra, but it does mean Theorem 1.2's upper containment is unproved as stated. I therefore recommend a conditional accept: either supply a correct proof of Proposition 5.2, or state the second inclusion with the missing |Im z| threshold. The reader's identified weakest assumption, the importation of (2.8)–(2.9), is a separate caveat; my concern is internal to the proof and was not the reader's weakest point.","tokens_in":4,"tokens_out":38039,"duration_ms":403974,"concrete_test":"Compute both sides of the claimed inequality for θ = π/4, m = 0, z = i/5, and ψ = (φ0,0,0,0)^T with φ0(x) = π^{-1/4}e^{-x²/2}. Since Aψ = 0, one has ||(A−i/5)ψ|| = 1/5, while ||Bψ|| = 2 sin(π/8)||xφ0|| = √2 sin(π/8) ≈ 0.541. The asserted bound would require 0.541 ≤ |tan(π/8)|·(1/5) ≈ 0.083, which is false. This directly disproves the proof step in Proposition 5.2. To settle the theorem's status, either repair this step with an additional hypothesis such as |Im z|² ≥ 2cos²(θ/2), or restrict the second inclusion of Theorem 1.2 accordingly.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The upper inclusion in Theorem 1.2 rests on Proposition 5.2, whose proof asserts that (3.2)–(3.3) imply ||Bψ|| ≤ |tan(θ/2)| ||(A−i Im z)ψ|| for all ψ. The exact identities are ||(A−i Im z)ψ||² = cos²(θ/2)(base+~) + (Im z)²||ψ||² and ||Bψ||² = sin²(θ/2)(base−~), where base = ||ψ'||²+||xψ||² and ~ = −||ψ1||²+||ψ2||²−||ψ3||²+||ψ4||². The desired inequality is equivalent to −2~ ≤ ((Im z)²/cos²(θ/2))||ψ||². For vectors with ~<0, e.g. ψ = (φ0,0,0,0)^T with φ0 the harmonic-oscillator ground state, ~ = −||ψ||², so the inequality requires |Im z|² ≥ 2cos²(θ/2). But the hypothesis of Proposition 5.2 with m = Re z = 0 only requires |Im z| > 0. Thus for θ = π/4 and z = i/5 the key inequality fails. Since this estimate is the sole basis for the second inclusion of Theorem 1.2, that containment is not established as written. Theorem 1.1 and the first inclusion of Theorem 1.2 do not depend on this step.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript introduces the rotated relativistic harmonic oscillator H_θ on L^2(R)^4, defines it as a closed operator with compact resolvent, and analyzes its spectrum, eigenfunctions, and pseudospectra. The main positive results are the explicit spectrum σ(H_θ) = {±√(2n+m^2)}, the construction of biorthonormal eigenfunctions via the supersymmetric square H_θ^2 = S_θ + m^2 + iα_1α_2, the exponential growth of spectral projectors in Theorem 1.1, and two pseudospectral inclusions in Theorem 1.2. The proofs import the known estimates (2.8) and (2.9) for the rotated Schrödinger operator S_θ.","tokens_in":16566,"tokens_out":9909,"duration_ms":93592,"significance":"If the main results stand, the paper gives a valuable explicit non-self-adjoint Dirac model: it provides an exact relativistic analogue of Davies' rotated oscillator, exhibits wild basis properties, and gives detailed pseudospectral information without any parameter fitting. The explicit ladder-operator construction and the projector asymptotics in Theorem 1.1 are particularly clean and appear correct. However, the upper pseudospectral inclusion in Theorem 1.2 rests on a false intermediate estimate, so the two-sided description of the pseudospectra and the claimed optimality discussion are not established as written.","major_comments":[{"comment":"The key estimate in the proof of Proposition 5.2 is not a consequence of (3.2)–(3.3). For k = Im z, the exact identities give ||(A − i Im z)ψ||² = cos²(θ/2)(base + τ) + (Im z)²||ψ||² and ||Bψ||² = sin²(θ/2)(base − τ), where base = ||ψ'||² + ||xψ||² and τ = −||ψ_1||² + ||ψ_2||² − ||ψ_3||² + ||ψ_4||². Comparing these shows that the claimed inequality ||Bψ|| ≤ |tan(θ/2)| ||(A − i Im z)ψ|| for all ψ requires (Im z)² ≥ 2 cos²(θ/2), not merely the angle condition stated in the proposition. The stated hypothesis permits arbitrarily small |Im z|: for instance, with m = 0, θ = π/4, z = i/5, and ψ = (φ_0, 0, 0, 0)^T where φ_0 is the normalized harmonic-oscillator ground state, one has ||Bψ||² = 2 sin²(θ/2) while |tan(θ/2)|² ||(A − i Im z)ψ||² = tan²(θ/2)(Im z)², so the inequality fails by a large margin. Therefore the application of Theorem 5.1 in Proposition 5.2 is invalid, and the second inclusion of Theorem 1.2 is not proved as written.","section":"5, Proposition 5.2"},{"comment":"The optimality discussion is directly dependent on Proposition 5.2. Since that proposition's proof is invalid, the claims that the second inclusion of Theorem 1.2 is optimal for small θ, and the associated discussion of the transition angle f(θ), are currently unsupported. The first inclusion of Theorem 1.2 and Theorem 1.1 are independent of this issue and are not affected.","section":"5, after Eq. (5.2)"}],"minor_comments":[{"comment":"The notation involving the symbol '~' is confusing: it appears both as an operator and as a scalar coefficient. Please define a named quantity, for example τ(ψ), so that the displayed identities are unambiguous.","section":"3.1, Eq. (3.2)–(3.3)"},{"comment":"The caption contains 'complex plain', which should be 'complex plane'.","section":"Figure 1 caption"},{"comment":"The keyword 'oscilator' is a typo for 'oscillator', and 'MCS 2020' should be 'MSC 2020'.","section":"Keywords and MSC"},{"comment":"Reference [30] contains 'and et al.' in the author list; this should be cleaned up before publication.","section":"References"},{"comment":"The use of concavity of the logarithm is fine, but the line would be easier to read if the quantities a = |u_1^1|² + |u_1^3|² and b = |u_1^2|² + |u_1^4|² were introduced explicitly before the inequality.","section":"4.3, Lemma 4.5"}],"recommendation":"major_revision","confidential_remarks":"The flaw in §5 is localized, and the rest of the paper is largely solid. The authors should be asked to repair or replace Proposition 5.2 and the second inclusion of Theorem 1.2; the current proof does not establish the upper pseudospectral bound, and the optimality claims should not be used until this is corrected. The paper is otherwise a good fit for the journal."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The main result—the exponential growth of the spectral projectors in Theorem 1.1—is new, correctly proved, and a worthwhile contribution to the non-self-adjoint Dirac literature. The construction of the model via H_theta^2 = S_theta + m^2 + i alpha1 alpha2 is transparent, and the non-relativistic limit is a nice addition. The lower pseudospectral inclusion in Theorem 1.2 follows cleanly from known results on S_theta and is not in question.\n\nThe soft spot is real and is exactly where the stress-test points. In Proposition 5.2 the authors claim the estimate ||Bpsi|| ≤ |tan(theta/2)| ||(A - i Im z)psi|| for all psi. That is false for small |Im z|. Take psi = (phi_0,0,0,0)^T with phi_0 the harmonic-oscillator ground state. Then A psi = 0 and ||B psi|| = sqrt(2) sin(theta/2) ||psi||, so the inequality would force |Im z| ≥ sqrt(2) cos(theta/2). The hypothesis of Proposition 5.2 with m = Re z = 0 only requires |Im z| > 0; for theta = pi/4 and z = i/5 the inequality crumbles. Since the second inclusion of Theorem 1.2 rests on this estimate, that containment is not established as written. The authors may still be right—numerically the pseudospectrum looks confined—but the proof needs a genuinely different argument, not a cosmetic repair. Theorem 1.1 and the first inclusion of Theorem 1.2 do not depend on this step and stand on their own.\n\nThe imported Schrödinger estimates (2.8)–(2.9) are another caveat, though not a flaw: the paper leans on established results about S_theta rather than reproving them, which is acceptable but leaves the Dirac results dependent on that external input. A self-contained treatment would have been stronger, though longer.\n\nWho gets value: readers working on non-self-adjoint Dirac operators, pseudospectra, and wild basis behaviour will want this model and Theorem 1.1. The paper deserves a serious referee, but not a quick accept: the author should be asked to fix Proposition 5.2 or downgrade the upper inclusion to a conjecture. I would take it to review with that expectation.","headline":"Solid new result on eigenprojector blow-up for a relativistic rotated oscillator, but the upper pseudospectral inclusion in Theorem 1.2 rests on a false estimate in Proposition 5.2 and needs repair.","tokens_in":17139,"tokens_out":22816,"would_cite":true,"duration_ms":174607,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["34L40","34L15","47A10","81Q12"],"pacs":[],"model":"deepseek-v4-flash","headline":"The relativistic rotated harmonic oscillator has a real discrete spectrum, but for any nonzero rotation its eigenfunctions fail to form a basis, with eigenprojector norms growing at the exact rate log sqrt((1+|sin θ|)/(1-|sin θ|)).","keywords":["pseudospectrum","resolvent estimate","relativistic harmonic oscillator","non-self-adjoint Dirac operator","rotated harmonic oscillator","Davies operator","wild basis properties"],"falsifier":"Compute the spectral projectors P_n^± numerically for a fixed rotation angle, say θ = π/4 and m = 1, for n up to a few hundred, and check whether log ||P_n^±||/n approaches log $\\sqrt$((1+sin θ)/(1-sin θ)) ≈ 0.4407; a systematic deviation from this value would refute Theorem 1.1. Alternatively, test the resolvent claim (1.10) by evaluating ||(H_θ ∓ m - r $e^{{iϑ}}$)^{-1}|| along the ray ϑ = θ/4 in the wedge (−θ/2, θ/2) as r grows: if the norms stay bounded instead of diverging, the pseudospectral picture fails.","tokens_in":16047,"feed_emoji":"⚛️","tokens_out":6227,"duration_ms":51652,"temperature":0.7,"pith_summary":"This paper constructs a relativistic counterpart of the complex-rotated harmonic oscillator, a Dirac operator on spinors, and shows that despite a real, discrete spectrum, its eigenfunctions behave wildly for any nonzero rotation angle. The central result pins the growth of the spectral projectors to the exact exponential rate log sqrt((1+|sin θ|)/(1-|sin θ|)), which forces the eigenfunctions to form neither a Riesz nor a Schauder basis. It further proves that the pseudospectra extend far from the spectrum in a wedge determined by the rotation angle, making the operator numerically pathological and not similar to a self-adjoint one. The interest is that this is an exactly solvable non-self-adjoint Dirac model with the same illusory simplicity as the non-relativistic rotated oscillator.","feed_headline":"Rotated Dirac oscillator: real spectrum, no eigenbasis","feed_subtitle":"Eigenprojector norms grow exponentially, so the Dirac oscillator can't be made self-adjoint by a bounded transform.","key_machinery":"The load-bearing object is the supersymmetric square identity $H_θ^{2}$ = (S_θ + $m^{2}$) I + i α1 α2, where S_θ = -$e^{{-iθ}}$ ∂$x^{2}$ + $e^{{iθ}}$ $x^{2}$ is the rotated Schrödinger oscillator; the extra matrix term is diagonal, so $H_θ^{2}$ decomposes into two copies of S_θ plus constants. This reduces the spectral analysis of the Dirac operator to the already understood non-relativistic operator S_θ. The proof of Theorem 1.1 then uses explicit biorthonormal eigenfunctions built from the rotated Hermite functions φ_n, with the norm of the eigenprojectors bounded in terms of ||φ_n||^2, whose exponential growth is quoted from the existing literature on S_θ. For the pseudospectral statements, the same identity transfers known pseudospectral inclusions for S_θ to H_θ.","core_discovery":"The paper's central claim is Theorem 1.1: for the relativistic rotated harmonic oscillator H_θ = -iα1 $e^{{-iθ/2}}$ ∂x - α2 $e^{{iθ/2}}$ x + m α3 on $L^{2}$(R)^4, the spectral projectors P_n^± onto the eigenspaces of ±√(2n+$m^{2}$) satisfy lim_{n→∞} (log ||P_n^±||)/n = log $\\sqrt$((1+|sin θ|)/(1-|sin θ|)). This is the exact Dirac analogue of the known growth for the rotated Schrödinger oscillator S_θ. As a consequence, whenever θ≠0 the eigenfunctions of H_θ form neither a Riesz nor a Schauder basis. The same supersymmetric identity $H_θ^{2}$ = (S_θ + $m^{2}$) I + i α1 α2 yields the spectrum σ(H_θ) = {±√(2n+$m^{2}$)} and, via known pseudospectral bounds for S_θ, a two-sided description of the ε-pseudospectra showing they are highly non-trivial.","pith_inferences":["A testable prediction left open by the paper is that for m = 0 the transition angle for pseudospectral growth is exactly θ/2, by analogy with the transition angle θ for S_θ; the numerical pseudospectra shown in the paper could be checked directly for this threshold.","The same supersymmetric transfer should work for other non-self-adjoint Dirac operators built from S_θ-type quadratic operators, for instance in higher dimensions or with external potentials, giving explicit exponential growth rates for their eigenprojectors.","If the exponential growth of eigenspace norms is generic for rotated relativistic oscillators, then any attempt to use such operators as quasi-Hermitian quantum models would need to confront this instability rather than assume a hidden self-adjointness."],"forward_implications":["For any θ ≠ 0, H_θ is not similar via a bounded and boundedly invertible transformation to a self-adjoint or normal operator, so the usual quantum-mechanical interpretation of observables fails for this model.","The eigenfunctions of H_θ do not form a Riesz or Schauder basis in L^2(R)^4, so expansions in eigenfunctions are unstable in a strong sense.","The ε-pseudospectrum of H_θ contains points arbitrarily far from the spectrum within the wedge |arg(z^2 - m^2)| ≤ θ - δ for any δ > 0, meaning numerical computation of eigenvalues and resonances is unstable.","Along rays inside the wedge (−θ/2, θ/2), the resolvent norm grows to infinity as the spectral parameter goes to infinity, as stated in equation (1.10).","In the non-relativistic limit c → ∞, the renormalized operator converges in norm-resolvent sense to a constant shift of the rotated Schrödinger oscillator S_θ, connecting the model to the well-known non-relativistic case."],"supporting_citations":[{"why":"supplies the quoted exponential growth rate (2.8) of the rotated Hermite functions and the pseudospectral inclusion (2.9) for S_θ on which Theorems 1.1 and 1.2 rest.","marker":"[26]"},{"why":"introduces the rotated harmonic oscillator S_θ whose relativistic analogue is the subject of the paper and whose wild pseudospectral behaviour motivates the whole analysis.","marker":"[12]"},{"why":"initiates the study of the non-self-adjoint harmonic oscillator and supports the pseudospectral claims about S_θ.","marker":"[4]"},{"why":"proves the complete pseudospectral description of the rotated harmonic oscillator, underwriting the inclusion used in Theorem 1.2.","marker":"[31]"},{"why":"provides the perturbation-theoretic bound used to define H_θ as a closed operator with compact resolvent and to prove the second inclusion of Theorem 1.2.","marker":"[22]"},{"why":"supplies the non-relativistic limit theorem used to show convergence of the renormalized Dirac operator to the rotated Schrödinger oscillator.","marker":"[33]"}],"fun_headline_variants":["Rotated Dirac oscillator: real spectrum, but no eigenbasis","Projector norms explode in rotated Dirac oscillator","Relativistic rotated oscillator: non-self-adjoint surprises","Real eigenvalues, yet no basis: rotated Dirac's catch"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole argument leans on the quoted estimate (2.8) that the norm of the rotated Hermite functions grows at the exponential rate log sqrt((1+|sin θ|)/(1-|sin θ|)), and on the pseudospectral inclusion (2.9) for S_θ; both are taken from the existing literature rather than reproved here, and if either is wrong or misquoted the central theorems fail.","fun_headline_variants_meta":{"raw":{"variants":["Rotated Dirac oscillator: real spectrum, but no eigenbasis","Projector norms explode in rotated Dirac oscillator","Relativistic rotated oscillator: non-self-adjoint surprises","Real eigenvalues, yet no basis: rotated Dirac's catch"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000199,"raw_usage":{"total_tokens":1300,"prompt_tokens":802,"completion_tokens":498,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":418,"completion_tokens_details":{"reasoning_tokens":432}},"tokens_in":418,"tokens_out":498,"duration_ms":5522,"temperature":1.0,"reasoning_tokens":432,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T13:06:38.081910+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the spectral projectors P_n^± numerically for a fixed rotation angle, say θ = π/4 and m = 1, for n up to a few hundred, and check whether log ||P_n^±||/n approaches log $\\sqrt$((1+sin θ)/(1-sin θ)) ≈ 0.4407; a systematic deviation from this value would refute Theorem 1.1. Alternatively, test the resolvent claim (1.10) by evaluating ||(H_θ ∓ m - r $e^{{iϑ}}$)^{-1}|| along the ray ϑ = θ/4 in the wedge (−θ/2, θ/2) as r grows: if the norms stay bounded instead of diverging, the pseudospectral picture fails.","supporting_citations":[{"cited_title":"Krejˇ ciˇ r ´ ık, P","cited_arxiv_id":null,"evidence_quote":"supplies the quoted exponential growth rate (2.8) of the rotated Hermite functions and the pseudospectral inclusion (2.9) for S_θ on which Theorems 1.1 and 1.2 rest."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"introduces the rotated harmonic oscillator S_θ whose relativistic analogue is the subject of the paper and whose wild pseudospectral behaviour motivates the whole analysis."},{"cited_title":"Boulton, The non-self-adjoint harmonic oscillator, compact semigr oups and pseudospectra , J","cited_arxiv_id":null,"evidence_quote":"initiates the study of the non-self-adjoint harmonic oscillator and supports the pseudospectral claims about S_θ."},{"cited_title":"Pravda-Starov, A complete study of the pseudo-spectrum for the rotated harm onic oscillator , J","cited_arxiv_id":null,"evidence_quote":"proves the complete pseudospectral description of the rotated harmonic oscillator, underwriting the inclusion used in Theorem 1.2."},{"cited_title":"Kato, Perturbation theory for linear operators , Springer-Verlag, Berlin, 1966","cited_arxiv_id":null,"evidence_quote":"provides the perturbation-theoretic bound used to define H_θ as a closed operator with compact resolvent and to prove the second inclusion of Theorem 1.2."},{"cited_title":"Thaller, The Dirac equation , Springer-Verlag, Berlin Heidelberg, 1992","cited_arxiv_id":null,"evidence_quote":"supplies the non-relativistic limit theorem used to show convergence of the renormalized Dirac operator to the rotated Schrödinger oscillator."}],"review_version":1}