{"id":"b9988362-3774-4886-af29-8b60452f2d24","arxiv_id":"2607.19067","paper_version":1,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"In the L^2-based metric completion induced by the BBM candidate metric, the BBM eigenfunctions are absent and the free transported Hamiltonian is the dilation generator with purely absolutely continuous spectrum.","lead":"This paper proves that the Bender-Brody-Müller Hamiltonian, a proposed non-Hermitian route to the Riemann zeros, fails inside its own natural Hilbert space completion: the candidate zero eigenfunctions do not live there. It also shows the free realization has purely continuous spectrum and that no bounded metric repair can make the BBM metric invertible.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"No significant objection identified.","rationale":"The reader's verdict ACCEPT is justified. I checked the key steps: the density of ΔD0 in L2 (Lemma S6), the unitary equivalence of the completion to L2 (Theorem S7), the spectral reduction of the free realization (Theorem S9), the deficiency-index family (Theorem S12), the real-point spectrum of the adjoint (Theorem S14), and the non-L2 tail of Δψ_z (Theorem S15). The bounded-sandwich no-go is proven by scaling the support, so only boundedness of h(D) is used. No circularity or missing proof was found. The sole caveat is the explicitly scoped choice to work with the standard L2-based completion; the paper repeatedly states this limitation, so it does not undermine the stated claim. The identity (27) is written only for x>1, but that is sufficient for non-membership because the divergence at infinity is already fatal.","tokens_in":12196,"tokens_out":29865,"duration_ms":274929,"concrete_test":"Verify that the obstruction is robust to the one modeling ambiguity on (0,1): recompute the proof of Theorem S15 with the Hurwitz-continuation shift convention Sψ_z(x)=ψ_z(x-1) on (0,1) instead of the hard-wall drop; since Δψ_z=x^{-z} then holds on all of R_+ and still has ||x^{-z}||_{L2(1,∞)}=∞ for Re z=1/2, the absence conclusion is unchanged.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central argument is internally consistent. The completion of C_c^∞(0,∞) in the η0 norm is unitarily L2 via Δ, the free transported realization is unitarily equivalent to the dilation generator, and Theorem S15's absence statement follows from Δψ_z=x^{-z} for x>1 (only the tail matters), so no self-adjoint realization in this completion can use the BBM functions as eigenvectors. The deficiency-index and adjoint-point-spectrum results in the Supplement check out; Theorem S4's non-invertibility of bounded sandwiches is a direct consequence of non-coercivity and needs only boundedness of h(D). The only caveat is the explicitly scoped choice of the L2-based completion: the paper does not claim to rule out rigged-Hilbert-space, distributional, or unbounded-metric formulations, and a different topology could in principle accommodate the BBM functions. This is a modeling assumption, not an internal flaw.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper analyzes the Hilbert-space completion induced by the Bender--Brody--Muller candidate metric \\eta_0=\\Delta^\\dagger\\Delta on the standard half-line core C_c^\\infty(0,\\infty). It shows that \\eta_0 is positive with trivial kernel but is not coercive, that the completion is canonically unitarily equivalent to L^2(\\mathbb R_+) via \\Delta, and that the free transported realization of the BBM formal expression is unitarily equivalent to the dilation generator, with purely absolutely continuous spectrum \\mathbb R. It further proves that no bounded sandwich \\Delta^\\dagger h(D)\\Delta is boundedly invertible; that the transported symmetric operator has deficiency indices (\\infty,\\infty) and its adjoint has every real point as an eigenvalue of infinite multiplicity; and that on the critical line \\Delta\\psi_z=x^{-z} is not in L^2, so the BBM eigenfunctions do not belong to the completion. The paper concludes that no self-adjoint realization in this L^2-based completion can realize the original BBM eigenfunction/boundary-condition mechanism for Riemann zeros, while explicitly excluding non-L^2, rigged-Hilbert-space, and distributional formulations from its scope.","tokens_in":12343,"tokens_out":40418,"duration_ms":306107,"significance":"If correct, this settles the status of the original BBM proposal within the standard L^2-based quasi-Hermitian framework: the metric topology alone excludes the candidate eigenfunctions, independently of the choice of self-adjoint extension. The paper provides a complete, internally consistent operator-domain analysis with explicit deficiency functions, a Mellin reduction, and a realization-independent obstruction. It also contributes standalone spectral facts, namely the absence of boundedly invertible metric sandwiches and the contrast between the adjoint's real point spectrum and the purely continuous free extension. The main limitation, which is properly acknowledged in the manuscript, is that the negative conclusion is relative to the L^2-based completion and does not classify rigged-Hilbert-space or distributional formulations. The supplemental proofs are detailed and, apart from the local issues noted below, check out.","major_comments":[],"minor_comments":[{"comment":"The proof of Lemma S6 states 'Thus ker \\Delta^\\dagger={0} and Ran \\Delta=L^2(\\mathbb R_+)'; the second equality is false, since ker \\Delta^\\dagger={0} only implies that Ran \\Delta is dense, and in fact Ran(I-S) is not all of L^2 (for example, the indicator 1_{(0,1)} is not in the range). The desired conclusion that \\Delta D_0 is dense still follows: Ran \\Delta is dense, D_0 is dense, and \\Delta is bounded, so \\Delta D_0 is dense in Ran \\Delta and hence in L^2. Please correct this intermediate statement.","section":"Supplement, Lemma S6"},{"comment":"The displayed definition '\\langle u,v\\rangle=\\int u(x)v(x)\\,dx, linear in the second argument' is not a positive-definite Hilbert inner product as written; the adjoint computations that follow use the convention \\langle u,v\\rangle=\\int \\overline{u(x)}v(x)\\,dx (linear in the second argument). Please fix the displayed definition.","section":"Supplement, inner-product convention"},{"comment":"The claim that a smooth function \\psi defines an element of \\mathcal H_{\\eta_0} only if \\Delta\\psi\\in L^2 is correct via the unitary identification of Theorem S7, but it would be helpful to state this explicitly rather than as a pointwise condition.","section":"Supplement, Theorem S15"}],"recommendation":"minor_revision","confidential_remarks":"I recommend minor revision. The central mathematical claim is sound, the proofs are detailed, and the paper is appropriately scoped; the required changes are local corrections to Lemma S6 and the inner-product definition."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The headline: this paper is a clean, honest negative result about the BBM Hamiltonian. It shows that if you take BBM's own candidate metric seriously and complete the standard half-line core in that metric, the completed space is just L^2(R+) via the map phi = Delta psi, the free realization is unitarily equivalent to the dilation generator, and the BBM eigenfunctions are not in the space because Delta psi_z = x^{-z}, which is not L^2 for Re z = 1/2. That kills the original BBM candidate-eigenfunction/boundary-condition mechanism for every self-adjoint extension in this completion. I think the claim is exactly as scoped and the argument holds.\n\nThe genuinely new pieces are the completion theorem, the bounded-sandwich no-go (no Delta-dagger h(D) Delta is boundedly invertible), and the deficiency-index/adjoint-point-spectrum facts. I verified the residue calculation for non-coercivity, the density of Delta D_0, the Mellin reduction, and the explicit deficiency functions; they are complete and internally consistent. No fitted parameters, no circularity. The citation pattern is fair: Bellissard, Moxley, and BBM's replies are credited with what they actually did, and the added content is precise.\n\nSoft spots: the scope is narrow by design. The no-go applies to the L^2-based completion only; rigged-Hilbert, distributional, or unbounded-metric formulations are explicitly left open. If someone thinks BBM's physical state space is one of those, the conclusion does not touch them. That is a modeling choice, flagged in the paper, not an oversight. The de Branges remark only rules out one natural E(z), and the paper says so. The adjoint point-spectrum result is interesting but irrelevant to the Riemann-zero question; it is a bonus, not a rescue.\n\nWho should read it: anyone working on pseudo-Hermitian quantum mechanics, the Hilbert-Polya program, or non-coercive metric completions. It deserves a serious referee; the supplement is written well enough that the referee work is mainly checking analytic details like the independence argument in Theorem S14. I would send it out, and I would expect accept after minor revision.","headline":"A clean, well-scoped negative result: BBM's eigenfunctions do not survive the metric completion, with a solid supplement and only the expected scope caveat.","tokens_in":12874,"tokens_out":2416,"would_cite":true,"duration_ms":21649,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["81Q12","47B25","11M26"],"pacs":["03.65.-w"],"model":"deepseek-v4-flash","headline":"The BBM Hamiltonian's own metric completion excludes its proposed zeta-zero eigenfunctions.","keywords":["BBM Hamiltonian","metric completion","pseudo-Hermitian quantum mechanics","dilation generator","non-coercive metric","deficiency indices","zeta-function zeros","purely continuous spectrum"],"falsifier":"Exhibit one self-adjoint extension of the transported symmetric operator $T$ in the $\\eta_0$-completion that has $\\psi_{1/2+i\\gamma}$ as an eigenvector with eigenvalue $-2\\gamma$, or equivalently find an $\\eta_0$-Cauchy sequence of test functions whose limit represents that function. The direct calculation to check is whether any element of the completion can have shift difference $x^{-z}$; since $\\int_1^\\infty x^{-1}\\,dx=\\infty$, such a representative cannot exist.","tokens_in":12004,"feed_emoji":"","tokens_out":10011,"duration_ms":79769,"temperature":0.7,"pith_summary":"This paper asks what Hilbert space the BBM Hamiltonian's candidate metric actually generates, and whether the proposed eigenfunctions survive that completion. It shows that completing the standard half-line test functions in the norm $\\|\\psi\\|_{\\eta_0}=\\|\\Delta\\psi\\|$ produces a Hilbert space canonically unitarily equivalent to $L^2(\\mathbb R_+)$, with the free transported Hamiltonian conjugate to the dilation generator and purely absolutely continuous spectrum $\\mathbb R$. The candidate eigenfunctions fail to survive: for $\\mathrm{Re}\\,z=1/2$ the shift difference $(\\Delta\\psi_z)(x)=x^{-z}$ is not square-integrable, so these functions are not vectors of the completed space. The paper concludes that no self-adjoint realization in this $L^2$-based metric completion can use the original BBM eigenfunctions and boundary condition to produce zeta-zero point-spectrum states.","feed_headline":"BBM metric completion cannot host its zeta-zero eigenstates","feed_subtitle":"The half-line completion of the candidate metric is L² with dilation spectrum ℝ, and the proposed ψ_z vectors are absent.","key_machinery":"The central object is the positive but non-coercive quadratic form $\\eta_0=\\Delta^\\dagger\\Delta=2I-\\hat S-\\hat S^\\dagger$, the half-line realization of BBM's proposed metric $\\sin^2(\\hat p/2)$. Its boundary symbol $\\eta_0(k)=2(1-\\cos k)$ vanishes at $k\\in2\\pi\\mathbb Z$, creating soft directions but no null vectors. The argument runs through the completion of $C_c^\\infty(0,\\infty)$ in the norm $\\|\\psi\\|_{\\eta_0}=\\|\\Delta\\psi\\|_{L^2}$; the shift difference $\\Delta$ extends to a unitary map onto $L^2(\\mathbb R_+)$, transporting the formal Hamiltonian to the dilation generator $D=-i(2x\\partial_x+1)$. The decisive identity is the shift-difference relation $(\\Delta\\psi_z)(x)=x^{-z}$, which puts the candidate eigenfunctions outside the completed space when $\\mathrm{Re}\\,z=1/2$. A separate mechanism is the no-go for bounded sandwiches: broad states $\\psi_n(x)=n^{-1/2}\\chi(x/n)$ have $\\|\\Delta\\psi_n\\|\\to0$, so $\\Delta^\\dagger h(D)\\Delta$ cannot be boundedly invertible for any bounded $h$.","core_discovery":"The central claim is that the BBM similarity structure fixes a metric topology that is fatal to the original Hilbert–Pólya mechanism, even though it supports many operator realizations. The positive form $\\eta_0=\\Delta^\\dagger\\Delta=2I-\\hat S-\\hat S^\\dagger$ is not coercive: normalized Cauchy-kernel probes concentrated near $k=2\\pi$ have metric length $\\langle\\phi_\\varepsilon,\\eta_0\\phi_\\varepsilon\\rangle=2(1-e^{-\\varepsilon})\\to0$. Completing $C_c^\\infty(0,\\infty)$ in $\\|\\Delta\\psi\\|$ gives a Hilbert space canonically unitarily equivalent to $L^2(\\mathbb R_+)$ via $\\Phi=\\Delta\\psi$; the free self-adjoint realization is the dilation generator $D$, whose spectrum is $\\mathbb R$, simple and purely absolutely continuous. Two statements hold independently of zeta: no bounded sandwich $\\Delta^\\dagger h(D)\\Delta$ is boundedly invertible, and the transported symmetric operator $T=D|_{\\Delta D_0}$ has deficiency indices $(\\infty,\\infty)$, with the adjoint having every real point as an eigenvalue of infinite multiplicity while the free extension is purely continuous. Because $(\\Delta\\psi_z)(x)=x^{-z}$ on $x>1$, the BBM eigenfunctions on the critical line lie outside the completion, so every self-adjoint extension in this space fails to realize the proposed zeta-zero eigenfunction/boundary-condition mechanism.","pith_inferences":["An implication left implicit is that the obstruction is vector-level rather than extension-level: choosing a different self-adjoint extension cannot restore the missing $\\psi_z$, so the BBM proposal's Hilbert-space formulation is not simply under-specified.","The sandwich no-go suggests that any repair of the BBM similarity must alter the shift factor $\\Delta$ itself or the factorization $\\Delta^{-1}D\\Delta$, not merely replace the metric by $\\Delta^\\dagger h(D)\\Delta$.","A testable extension is to ask whether an unbounded metric of the form $\\Delta^\\dagger h(D)\\Delta$, with $h$ unbounded, can define a Hilbert topology that contains the critical-line functions $x^{-1/2-i\\gamma}$; the paper neither constructs nor excludes such a space.","The deficiency-index phenomenon, with infinitely many extensions, a purely continuous free realization, and point spectrum in the adjoint, is likely generic for formal Hamiltonians built from a non-coercive similarity and could be probed in other shift-difference models."],"forward_implications":["In the $\\eta_0$-completion, the free transported BBM Hamiltonian is unitarily equivalent to the dilation generator and has purely absolutely continuous spectrum $\\mathbb R$, with no isolated levels.","The formal BBM operator has infinitely many self-adjoint extensions, so the formal expression alone does not select a physical Hamiltonian.","No bounded-multiplier repair of the metric can restore coercivity, ruling out a whole class of attempted pseudo-Hermitian fixes.","If point spectrum exists in this completion, it must come from eigenvectors other than the original BBM functions $\\psi_z$.","Any BBM-type construction that hopes to recover zeta zeros as point spectrum must leave the $L^2$-based completion, for example by adopting a non-$L^2$, rigged, distributional, or unbounded-metric state space."],"supporting_citations":[{"why":"Supplies the BBM Hamiltonian, the candidate metric $\\sin^2(\\hat p/2)$, the eigenfunctions $\\psi_z$, and the boundary condition $\\psi_z(0)=0$ that the paper tests.","marker":"[1]"},{"why":"Shows the half-line momentum operator has no self-adjoint extension and the BBM eigenfunctions are not $L^2$ on the critical line, the backdrop for the completion analysis.","marker":"[2]"},{"why":"Introduces the formal conjugation of the BBM eigenvalue equation to a dilation equation, the similarity structure used here.","marker":"[3]"},{"why":"States the BBM authors' pseudo-Hermitian/biorthogonal reply, whose metric-topology claim the paper examines.","marker":"[4]"},{"why":"Supplies the $H=xp$ dilation-generator spectral picture to which the free realization is compared.","marker":"[5]"},{"why":"Gives the pseudo-Hermitian metric condition $\\eta H=H^\\dagger\\eta$ that frames BBM's physical interpretation.","marker":"[7]"},{"why":"Provides the standard quasi-Hermitian $L^2$-based completion construction that the paper applies to the BBM metric.","marker":"[14]"},{"why":"Early exactly solvable non-Hermitian model in which the metric is not equivalent to the ambient topology, used as a precedent for the same obstruction.","marker":"[16]"}],"fun_headline_variants":["BBM completion: no zeta zero eigenstates","Dilation spectrum, zeta states absent in BBM","No zeta states in BBM metric completion","BBM completion: dilation spectrum, no eigenstates"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the physical state space is the Hilbert space obtained by completing smooth compactly supported half-line functions in the $\\eta_0$ norm; if one instead adopts a more permissive topology such as a rigged Hilbert space, a distributional space, or an unbounded-metric formulation, the BBM eigenfunctions could in principle survive.","fun_headline_variants_meta":{"raw":{"variants":["BBM completion: no zeta zero eigenstates","Dilation spectrum, zeta states absent in BBM","No zeta states in BBM metric completion","BBM completion: dilation spectrum, no eigenstates"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00073,"raw_usage":{"total_tokens":3369,"prompt_tokens":1146,"completion_tokens":2223,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":762,"completion_tokens_details":{"reasoning_tokens":2160}},"tokens_in":762,"tokens_out":2223,"duration_ms":14160,"temperature":1.0,"reasoning_tokens":2160,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T15:31:44.885791+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Exhibit one self-adjoint extension of the transported symmetric operator $T$ in the $\\eta_0$-completion that has $\\psi_{1/2+i\\gamma}$ as an eigenvector with eigenvalue $-2\\gamma$, or equivalently find an $\\eta_0$-Cauchy sequence of test functions whose limit represents that function. The direct calculation to check is whether any element of the completion can have shift difference $x^{-z}$; since $\\int_1^\\infty x^{-1}\\,dx=\\infty$, such a representative cannot exist.","supporting_citations":[{"cited_title":"This is stronger than saying that a formal differential expression has a domain prob- lem","cited_arxiv_id":null,"evidence_quote":"Supplies the BBM Hamiltonian, the candidate metric $\\sin^2(\\hat p/2)$, the eigenfunctions $\\psi_z$, and the boundary condition $\\psi_z(0)=0$ that the paper tests."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Shows the half-line momentum operator has no self-adjoint extension and the BBM eigenfunctions are not $L^2$ on the critical line, the backdrop for the completion analysis."},{"cited_title":"Connes,Trace formula in noncommutative geometry and the zeros of the Riemann zeta function, Selecta Math","cited_arxiv_id":null,"evidence_quote":"Gives the pseudo-Hermitian metric condition $\\eta H=H^\\dagger\\eta$ that frames BBM's physical interpretation."},{"cited_title":"Mostafazadeh,Pseudo-Hermitian representation of quantum mechanics, Int","cited_arxiv_id":null,"evidence_quote":"Early exactly solvable non-Hermitian model in which the metric is not equivalent to the ambient topology, used as a precedent for the same obstruction."}],"review_version":2}