{"id":"fa9cdd30-9acf-4dcb-9d55-9b869f27eb95","arxiv_id":"2411.12501","paper_version":2,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"The imaginary cubic oscillator is reinterpreted as an unphysical intrinsic exceptional point that can only be recovered as a singular limit of suitably perturbed Hamiltonians.","lead":"An old quantum toy model, the imaginary cubic oscillator, is argued to be physically uninterpretable as a true Hamiltonian, only a singular limit of a regularized family. The paper develops a perturbation framework for understanding such 'intrinsic exceptional points' in quantum theory.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The central claim depends on R(IEP) being a genuine basis and on H(new)(lambda) actually existing; neither is demonstrated, and Sec. 7.1 concedes that the latter is open.","rationale":"The reader's weakest_assumption identifies both missing ingredients: the legitimacy of the auxiliary vectors produced by recurrences (28)–(30) and the existence of a physically admissible one-parameter family H(new)(lambda). My stress-test agrees that these are the load-bearing points. The paper is transparent about its programmatic status—it calls H(new)(lambda) 'hypothetical' and admits in Sec. 7.1 that identifying benign perturbations is an open problem. But that transparency does not reduce the gap: the central claim is an interpretive statement about what H(IC) 'can only' be, and the only constructive evidence offered is an analogy with finite-dimensional EPN unfolding. The finite-N case is not dispositive at N=infty, since the very result being addressed is the failure of the Riesz-basis property in the infinite-dimensional limit. A revision that explicitly constructs H(new)(lambda) for H(IC), or proves that the proposed R(IEP) is a bounded operator with bounded inverse, would change the assessment. Until then, the CONDITIONAL verdict is appropriate. The concrete test I propose directly probes the basis construction; a numerical divergence of the norms or Gram condition number would show that the regularization step fails, while a convergent result would strengthen the paper's programmatic case. I therefore recommend no change to the reader's verdict.","tokens_in":24496,"tokens_out":7110,"duration_ms":72747,"concrete_test":"Discretize H(IC) (e.g., in a 400-, 800-, and 1600-dimensional spectral basis), compute the lowest eigenpairs, fix K=50, and solve recurrence (28)–(30) for |f_{K+k}>, k=1,...,N-K, using the stated freedom in c_{k,k} (try c_{k,k}=1 and try normalizing ||f_{K+k}||=1). Compute ||f_{K+k}|| and the condition number of the truncated Gram matrix G_{ij}=<f_i|f_j> as N grows. If either grows without bound, R(IEP) is not a bounded operator with bounded inverse on H, so Eq. (25) and the perturbation expansion (31) are not justified and the central claim is unsupported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The abstract's central claim—that H(IC) can only be read as the unphysical IEP limit of a one-parameter family of standard Hamiltonians—is not demonstrated. The constructive core in Sec. 5.3 defines the replacement vectors |f_{K+p}> by the finite sums (29), with coefficients fixed by recurrence (30). Although each |f_{K+p}> is formally a finite linear combination of eigenstates, no argument is given that these vectors have finite, controllable L2 norms as p grows, that the infinite set (27) spans H, or that the resulting R(IEP) has a bounded inverse; without those properties Eq. (25) is only a formal identity and the perturbed equation (31) has no operator meaning. This is not a technicality: the Siegl–Krejcírík obstruction is precisely the failure of the eigenbasis to be a Riesz basis, and the proposed fix must be shown to restore the Riesz property. In parallel, Sec. 6.1 invokes a 'hypothetical' family H(new)(lambda) with H(new)(0)=H(IC), and Sec. 7.1 states that identifying benign, unitarity-preserving perturbations is 'a mathematically much more difficult open problem.' Thus the two objects on which the central claim rests—a well-behaved R(IEP) and an actual family H(new)(lambda)—are assumed rather than established. The finite-N EPN analogy of Secs. 2–3 motivates the construction but cannot substitute for a proof at N=infty, especially because the paper's own point is that the infinite-dimensional case is essentially different.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper revisits the imaginary cubic oscillator Hamiltonian H^(IC)=p^2+i x^3 and the Siegl–Krejčířík result that its eigenvectors do not form a Riesz basis. It proposes that H^(IC) is an 'intrinsic exceptional point' (IEP) at N=∞, analogous in several respects to a finite-dimensional Jordan-block exceptional point, and that it should be regarded as a singular, unphysical limit of a one-parameter family of standard quasi-Hermitian Hamiltonians. The constructive part defines a replacement transition matrix R^(IEP) whose columns are the first K eigenvectors plus associated vectors |f_{K+p}> obtained by the recurrences (28)–(30), and it sketches a perturbation theory around the canonical form J^(IEP)+λV. The paper concludes that unitarity can only be restored by perturbing away from the IEP and that the bare IC Hamiltonian is not a legitimate closed-system observable.","tokens_in":24828,"tokens_out":5532,"duration_ms":55670,"significance":"If the main claim could be made rigorous, it would settle a longstanding interpretive question about the imaginary cubic oscillator and would have consequences for the wider class of PT-symmetric models with non-Riesz eigenbases. The paper is valuable as a conceptual framework: it develops the finite-N EPN analogy, isolates the relevant technical difficulties, and is explicit about where the argument is incomplete, especially in Sec. 7.1. It does not, however, currently supply the functional-analytic proof needed to turn the conjecture into a theorem: there are no norm estimates or Riesz-basis checks for the proposed replacement basis, no explicit admissible perturbation family for H^(IC), and no convergence proof for the perturbation series. The manuscript after a substantial revision could become a useful contribution, but as it stands the central assertion is not demonstrated.","major_comments":[{"comment":"The construction of the associated vectors |f_{K+p}> is purely formal. The text defines these vectors as finite sums over eigenstates, but it gives no estimate of their norms, no proof that the infinite set in Eq. (27) is complete in the Hilbert space, and no proof that the transition operator R^(IEP) has a bounded inverse. Without these properties Eq. (25) is only a formal identity, and Eq. (31) is not a legitimate similarity transformation on the Hilbert space. This is not a technicality: the Siegl–Krejčířík obstruction is precisely the failure of the eigenbasis to be a Riesz basis, so the proposed replacement basis must be shown to restore the Riesz property. The sentence 'The goal is achieved' following Eq. (30) therefore overstates what has been shown.","section":"Sec. 5.3, Eqs. (28)–(30)"},{"comment":"The central claim of the abstract—that H^(IC) 'can only be interpreted' as the singular IEP limit of a one-parameter family of standard Hamiltonians—rests on the existence of a physically admissible family H^(new)(λ) with H^(new)(0)=H^(IEP). No admissible family is actually constructed for the imaginary cubic oscillator; Eq. (31) provides a template for one, but its admissibility requires properties of R^(IEP) and V that are not established. The paper itself calls the family 'hypothetical' and states in Sec. 7.1 that identifying benign, unitarity-preserving perturbations is 'a mathematically much more difficult open problem.' The finite-N EPN analogy of Secs. 2–3 motivates the idea but cannot establish the N=∞ statement, especially since Sec. 4.1 emphasizes that the analogy is incomplete. As written, the abstract's conclusion should be presented as a conjecture or research program rather than an established result.","section":"Sec. 6.1, Eq. (31), and Sec. 7.1"},{"comment":"The perturbation series for |ψ(λ)> and E(λ) are assumed to converge, and the triangular-matrix inversion leading to Eq. (41) is purely algebraic. In the infinite-dimensional Q-projected sector no convergence or domain argument is provided. Since the unperturbed operator is unbounded and non-normal, term-by-term operations require justification; otherwise the leading-order criteria developed here remain formal. A concrete check would be to exhibit a perturbation V for which the series have a nonzero radius of convergence and satisfy the relevant estimates, at least for a truncated but N-independent version of the problem.","section":"Sec. 6.2, Eqs. (33)–(34)"}],"minor_comments":[{"comment":"The matrix element typeset as 'V201' in the second row of the perturbation matrix should presumably be 'V20'.","section":"Eq. (39)"},{"comment":"The second-order energy correction contains a stray ket symbol: 'E[2]⟩' should be 'E[2]'.","section":"Eq. (34)"},{"comment":"The word 'constructiton' should be 'construction'.","section":"Sec. 6.2, opening paragraph"},{"comment":"The approximation '≈' in the asymptotic parallelization statements is not quantified; stating in which norm and at what rate the eigenvectors become parallel would make the IEP intuition testable.","section":"Sec. 4.1, Eqs. (23)–(24)"},{"comment":"The displayed matrix J^(IEP) is difficult to parse; a block-matrix presentation with explicit P and Q subblocks would improve readability.","section":"Eq. (26)"}],"recommendation":"major_revision","confidential_remarks":"The paper draws heavily on the author's own prior work (Refs. [11,17,18,20]), but those works are used as templates rather than as fitted inputs, so I do not see a novelty or disclosure problem. The main risk is that the central claim is presented as a conclusion while the manuscript itself acknowledges the decisive open problem in Sec. 7.1. The authors should be given the opportunity either to supply the missing analytic control or to reframe the contribution as a conjecture supported by the finite-N analogy."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's the thing you need to know: Znojil's new paper is honest but programmatic. The genuinely new piece is the construction of an 'IEP-unfolding basis'—Eq. (27) with the |f> vectors defined by recurrences (28)-(30)—which transfers his earlier EPN perturbation machinery to the infinite-dimensional intrinsic exceptional point setting. That's a reasonable formal move, and the paper earns credit for laying the EPN-IEP analogy out clearly and for being upfront that the Jordan-like structure (26) is a proposal rather than a derived result.\n\nThe soft spot is the central claim. The abstract says H^(IC) 'can only be interpreted as a manifestly unphysical, singular IEP limit of a hypothetical one-parametric family of certain standard quantum Hamiltonians.' The word 'hypothetical' is doing a lot of work. No such family is exhibited for the imaginary cubic oscillator, and Sec. 7.1 concedes that identifying benign, unitarity-preserving perturbations is 'a mathematically much more difficult open problem.' That admission matches the structure of the argument: Eq. (25) defines R(IEP) only formally, and the stress-test note is right that the paper doesn't show the |f> vectors have finite, controllable L2 norms, that the set (27) spans the space, or that R(IEP) has a bounded inverse. Without those, Eq. (31) is a formal identity, not an operator statement. So the central interpretive punchline—that the IC oscillator is only physical as a singular limit—is a well-motivated conjecture, not a demonstrated theorem.\n\nI want to balance that: the paper is not pretending otherwise. It repeatedly flags the hypothetical status, and the recurrences are concrete enough that someone could try to make them rigorous. That's why I wouldn't reject it out of hand. But a referee should push hard on exactly those two points: either construct an explicit H^(new)(λ) for the IC model, or prove that the proposed basis actually restores the Riesz property. Without one of those, the abstract overclaims relative to the results.\n\nWho should read this: specialists in PT-symmetric and quasi-Hermitian quantum mechanics, especially anyone still using the imaginary cubic oscillator as a benchmark. It's a useful programmatic contribution, but it is not the last word. I'd send it to peer review with the expectation of major revision.","headline":"An honest programmatic paper that transfers EPN perturbation tricks to the IEP setting, but the central interpretive claim remains a conjecture rather than a proof.","tokens_in":25339,"tokens_out":3100,"would_cite":true,"duration_ms":28788,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["81Q12","81Q10","47A55"],"pacs":["03.65.-w"],"model":"deepseek-v4-flash","headline":"The popular imaginary cubic oscillator $p^2+ix^3$, despite its real spectrum and unbroken PT-symmetry, cannot serve as a closed-system quantum Hamiltonian; the paper argues it is only an unphysical \"intrinsic exceptional point\" limit of a…","keywords":["intrinsic exceptional point","imaginary cubic oscillator","PT-symmetric quantum mechanics","quasi-Hermitian Hamiltonian","Riesz basis","perturbation theory","exceptional point","non-Hermitian quantum mechanics"],"falsifier":"Compute the asymptotic overlap of consecutive high-energy eigenvectors of $H^{(IC)}$: if $|\\langle\\psi_{M+k}|\\psi_{M+k+1}\\rangle|$ stays bounded away from 1 as $M\\to\\infty$, the parallelization that drives the argument is absent. Alternatively, exhibit one explicit one-parameter family $H^{(new)}(\\lambda)$ with $H^{(new)}(0)=H^{(IC)}$ whose eigenvectors form a well-behaved basis for small $\\lambda\\neq 0$, which would supply the regularized Hamiltonian the paper says is only hypothetical.","tokens_in":24188,"feed_emoji":"⚛️","tokens_out":19723,"duration_ms":166992,"temperature":0.7,"pith_summary":"The paper argues that the imaginary cubic oscillator, $H^{(IC)}=p^2+ix^3$, despite having a real, discrete, bounded-below spectrum and unbroken PT-symmetry, cannot be accepted as a Hamiltonian of a closed quantum system. The obstruction is an \"intrinsic exceptional point\" (IEP): at high energies the eigenvectors become asymptotically parallel, so they do not form a well-behaved basis and no physical inner-product metric can be constructed. The paper proposes that $H^{(IC)}$ and similar operators be read only as the singular, unphysical IEP limit of a hypothetical one-parameter family of standard Hamiltonians. If this is right, the model's long-standing role as a benchmark of PT-symmetric quantum mechanics must be replaced by a perturbation-regularized picture in which unitarity is restored only away from the singular point.","feed_headline":"Imaginary cubic oscillator fails as a quantum Hamiltonian","feed_subtitle":"Real spectrum and PT symmetry are not enough when high-energy eigenvectors stop forming a usable basis.","key_machinery":"The load-bearing object is the infinite transition matrix $R^{(IEP)}=\\{|\\psi_0\\rangle,\\dots,|\\psi_{K-1}\\rangle,|f_K\\rangle,|f_{K+1}\\rangle,\\dots\\}$, the analogue of the transition matrix used at finite-order exceptional points. It is defined to satisfy $H^{(IEP)}R^{(IEP)}=R^{(IEP)}J^{(IEP)}$, with $J^{(IEP)}$ block-diagonal: the first $K$ eigenstates keep their energy eigenvalues, while the asymptotic tail becomes a two-diagonal matrix with $E_{K+m}$ on the diagonal and 1 on the superdiagonal. The columns $|f_{K+m}\\rangle$ are generated recursively by $(H^{(IEP)}-E_{K+m})|f_{K+m}\\rangle=|f_{K+m-1}\\rangle$ with $|f_K\\rangle=c_{0,0}|\\psi_K\\rangle$, which in the eigenbasis yields the coefficient recurrence $c_{k,m}=(E_{K+m}-E_{K+k})^{-1}c_{k-1,m}$. This construction is what removes the asymptotic parallelization and makes an amended perturbation theory of the form $[R^{(IEP)}]^{-1}H^{(new)}(\\lambda)R^{(IEP)}=J^{(IEP)}+\\lambda V$ possible.","core_discovery":"On the paper's own terms, the discovery is that the failure of the imaginary cubic oscillator is not an accidental spectral pathology but a new type of singularity, the intrinsic exceptional point, which stands to an ordinary exceptional point of finite order $N$ as an $N=\\infty$ limit stands to the finite case. The spectrum remains non-degenerate, yet the right eigenvectors and the left eigenvectors separately undergo asymptotic parallelization at high energies, so the eigenbasis is complete but not unconditional. Replacing the high-energy eigenvectors by non-eigenvector columns built from the recurrence $(H^{(IEP)}-E_{K+m})|f_{K+m}\\rangle=|f_{K+m-1}\\rangle$ de-parallelizes the basis and yields a transition matrix $R^{(IEP)}$ satisfying $H^{(IEP)}R^{(IEP)}=R^{(IEP)}J^{(IEP)}$, where $J^{(IEP)}$ is diagonal in the low-lying subspace and two-diagonal in the tail. The paper's central conclusion is that the IEP operator itself is manifestly unphysical and can only be interpreted as the singular limit of a hypothetical one-parameter family of standard Hamiltonians; only a perturbation away from the point can restore unitarity and physicality.","pith_inferences":["If the paper's claim is right, numerically computed bound states of the bare $H^{(IC)}$ should be treated with suspicion: finite-dimensional truncations quietly regularize the singularity, so results can depend on the cutoff in a way that masks the unphysicality of the exact operator.","The paper's distinction between benign and malign perturbations suggests a concrete research program: classify small perturbations by whether the perturbed spectrum stays real and the eigenvectors retain a well-behaved basis, using the leading-order criteria of Section 7.1 as a first filter.","One testable extension would be to compute the pseudospectra of $H^{(IC)}$ at high energies: if the asymptotic parallelization is real, the spectral instability should grow dramatically with energy, giving numerical evidence that is accessible before the mathematical basis question is settled."],"forward_implications":["The imaginary cubic oscillator cannot serve as the Hamiltonian of a closed, unitary quantum system, so quantum-mechanical predictions made directly from $H^{(IC)}$ lack a consistent probabilistic interpretation.","Any physically acceptable model sharing the IC spectrum must be a regularized perturbation of the IEP seed, with unitarity restored only when the perturbation parameter moves the operator away from the singularity.","Perturbation theory for such models should be built in the transition basis $R^{(IEP)}$, starting from the non-diagonal $J^{(IEP)}$, rather than in the eigenbasis of $H^{(IEP)}$.","The low-lying states can be treated by ordinary textbook perturbation theory once a sufficiently large cutoff $K$ is chosen, while the high-lying asymptotic states require the amended recurrences.","The same perturbation-regularization reading extends to other non-Hermitian models whose asymptotic eigenvectors parallelize, i.e., to IEP models with $N=\\infty$."],"supporting_citations":[{"why":"proves the imaginary cubic oscillator has no metric operator and no Riesz basis of eigenvectors, the mathematical fact the paper reinterprets as an intrinsic exceptional point.","marker":"[1]"},{"why":"introduced the PT-symmetric Hamiltonian class and the spectral-reality conjecture that made the imaginary cubic oscillator a benchmark model.","marker":"[5]"},{"why":"defines ordinary exceptional points and the perturbation-theory setting whose finite-N machinery is transferred to the IEP case.","marker":"[7]"},{"why":"supplies independent numerical and analytic evidence for the tendency of the IC eigenvectors toward collinearity and isotropy at spectral infinity.","marker":"[8]"},{"why":"provides exactly solvable quasi-Hermitian models whose passage through an exceptional point is given a physical interpretation, the template for reading IEPs as limits.","marker":"[11]"},{"why":"gives the amended perturbation-theoretic construction and smallness criteria near finite exceptional points that the paper adapts to the IEP setting.","marker":"[17]"},{"why":"derives the condition for unitarity-preserving perturbations near an exceptional point, used in Section 7.1 to distinguish benign from malign perturbations.","marker":"[18]"}],"fun_headline_variants":["PT-symmetric oscillator hits intrinsic exceptional point","Quantum failure: imaginary cubic oscillator's hidden singularity","Intrinsic exceptional point makes cubic oscillator unphysical","Cubic oscillator's asymptotic parallelization breaks quantum rules","Beyond exceptional points: intrinsic singularity in PT-symmetric model"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument depends on assuming that the auxiliary vectors built from the oscillator's eigenvectors are genuine, well-behaved members of the Hilbert space that form a usable basis, and that a one-parameter family of ordinary Hamiltonians passing through the imaginary cubic oscillator actually exists.","fun_headline_variants_meta":{"raw":{"variants":["PT-symmetric oscillator hits intrinsic exceptional point","Quantum failure: imaginary cubic oscillator's hidden singularity","Intrinsic exceptional point makes cubic oscillator unphysical","Cubic oscillator's asymptotic parallelization breaks quantum rules","Beyond exceptional points: intrinsic singularity in PT-symmetric model"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000239,"raw_usage":{"total_tokens":1511,"prompt_tokens":936,"completion_tokens":575,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":552,"completion_tokens_details":{"reasoning_tokens":502}},"tokens_in":552,"tokens_out":575,"duration_ms":5027,"temperature":1.0,"reasoning_tokens":502,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T17:27:19.399193+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the asymptotic overlap of consecutive high-energy eigenvectors of $H^{(IC)}$: if $|\\langle\\psi_{M+k}|\\psi_{M+k+1}\\rangle|$ stays bounded away from 1 as $M\\to\\infty$, the parallelization that drives the argument is absent. Alternatively, exhibit one explicit one-parameter family $H^{(new)}(\\lambda)$ with $H^{(new)}(0)=H^{(IC)}$ whose eigenvectors form a well-behaved basis for small $\\lambda\\neq 0$, which would supply the regularized Hamiltonian the paper says is only hypothetical.","supporting_citations":[{"cited_title":"On the metric operator for the imaginary cubic oscillator","cited_arxiv_id":null,"evidence_quote":"proves the imaginary cubic oscillator has no metric operator and no Riesz basis of eigenvectors, the mathematical fact the paper reinterprets as an intrinsic exceptional point."},{"cited_title":"M., Boettcher, S","cited_arxiv_id":null,"evidence_quote":"introduced the PT-symmetric Hamiltonian class and the spectral-reality conjecture that made the imaginary cubic oscillator a benchmark model."},{"cited_title":"Perturbation Theory for Linear Operators ; Springer: Berlin, Germany, 1966","cited_arxiv_id":null,"evidence_quote":"defines ordinary exceptional points and the perturbation-theory setting whose finite-N machinery is transferred to the IEP case."},{"cited_title":"Passage through exceptional point: Case study","cited_arxiv_id":null,"evidence_quote":"provides exactly solvable quasi-Hermitian models whose passage through an exceptional point is given a physical interpretation, the template for reading IEPs as limits."},{"cited_title":"Admissible perturbations and false instabilities in PT- symmetric quantum systems","cited_arxiv_id":null,"evidence_quote":"gives the amended perturbation-theoretic construction and smallness criteria near finite exceptional points that the paper adapts to the IEP setting."},{"cited_title":"Znojil, M","cited_arxiv_id":null,"evidence_quote":"derives the condition for unitarity-preserving perturbations near an exceptional point, used in Section 7.1 to distinguish benign from malign perturbations."}],"review_version":1}