{"id":"7be0b310-a887-48b5-b520-6a1e4d14e25d","arxiv_id":"1908.03699","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Restricting linear Schrödinger evolution to invariant families such as Gaussian and coherent states yields nonlinear, sometimes linear, reduced dynamics; a pulled-back Lagrangian formalism extends the construction to non-invariant families as an approximation.","lead":"This paper shows that a linear quantum equation, the Schrödinger equation, can produce nonlinear dynamics when it is restricted to a specially chosen family of states, such as Gaussian wave packets. It also gives a Lagrangian method to obtain approximate, finite-dimensional dynamics even when the chosen family is not preserved by the evolution.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The advertised extension to non-invariant submanifolds is untested: the anharmonic-oscillator equations (4.60) are never compared with the exact Schrödinger flow, so the claim that the pulled-back Lagrangian gives an approximation of the dynamics is unsupported.","rationale":"The reader's weakest assumption matches my own: the non-invariant-submanifold construction is the central new step, and its accuracy is unexamined. I agree with that assessment. The invariant-submanifold examples are self-contained and support the nonlinearity mechanism: the Gaussian calculation (2.6)–(2.15) and the harmonic version (2.35)–(2.37) explicitly verify that a linear unitary flow induces nonlinear equations on the parameter manifold, and the Lagrangian pullback in §4.1–4.2 gives the same equations, so the coincidence claim is credible in those cases. The gap is the anharmonic example: (4.60) is a concrete finite-dimensional system, but no evidence connects it to the true evolution. The abstract and introduction explicitly promise an approximation generalizing the variational method, so this missing comparison is not a peripheral omission; it is the part of the paper that would justify the title for non-invariant submanifolds. I also note a local sign inconsistency in the Radcliffe example, but the main reason to keep the verdict conditional is that the novel approximation claim is untested. The paper is not internally inconsistent, and the invariant-submanifold results are valuable, so rejection would be too strong. The conditional verdict should stand: accept after the anharmonic reduced dynamics is compared with the exact Schrödinger flow, or after an analytical error bound is supplied.","tokens_in":23745,"tokens_out":9224,"duration_ms":102476,"concrete_test":"Integrate the full Schrödinger equation for the anharmonic Hamiltonian (4.57) with, for example, ω = 1, λ = 0.1, and an initial Gaussian state with a_R = 1, a_I = 0, b = 0. At each time t, project ψ_exact(t) onto the Gaussian submanifold by maximizing |⟨ψ_g(a,b,c)|ψ_exact(t)⟩| over (a,b), and compare the projected (a_R(t), a_I(t), b_R(t), b_I(t)) with the solution of (4.60) on a fixed interval t ∈ [0,T]. If the mismatch is not O(λ) uniformly as λ → 0 on that interval, or if the projected exact trajectory and the reduced trajectory diverge at times where the λx^4 term is non-negligible, then the advertised approximation claim for non-invariant submanifolds fails.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The structural results for invariant submanifolds are sound: restricting the Schrödinger vector field to the invariant Gaussian submanifold yields the nonlinear systems (2.15) and (2.37), and the Lagrangian pullback reproduces them in §4.1–4.2 after the lift (4.29). The load-bearing gap is the advertised extension to non-invariant submanifolds. In §4, the paper defines Γ~ by i_{Γ~}ω~ = -dE~ (Eqs. 1.8 and 4.22) and claims this gives an approximation of the unitary flow φ_t, generalizing the variational method. The only nontrivial test is the anharmonic oscillator: Eqs. (4.60) are derived, but the text states that \"an extensive comparison between solutions of the equations of the motion on the submanifold of Gaussian states and their quantum evolution is beyond the scope of this paper and will be addressed elsewhere,\" and the conclusion repeats that approximation procedures are left to future work. No error estimate, no numerical comparison, and no theorem bounding the deviation from the exact flow is provided. For a non-invariant submanifold, a pulled-back Lagrangian always defines some dynamical system on M, but without additional control there is no reason that its trajectories stay close to φ_t or to the projection of φ_t onto the submanifold. Since this approximating property is exactly the claim being advertised, the central novel assertion is unsupported. A sign inconsistency in the Radcliffe example (4.15)–(4.17) also indicates that the examples need care, but the missing fidelity test is the principal concern.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies how a linear unitary flow on a Hilbert space induces dynamics on submanifolds of states. For invariant submanifolds (Gaussian states, coherent-state families), the authors compute the reduced nonlinear equations for free and harmonic-oscillator Hamiltonians and for spin/bosonic/fermionic coherent states. For non-invariant submanifolds, they propose a covariant pullback of a degenerate Lagrangian for the Schrödinger equation, Eq. (1.8), which yields a reduced Hamiltonian system on the parameter manifold; the anharmonic oscillator is treated as the example. The invariant-submanifold part is explicit and mostly self-contained; the non-invariant approximation claim is not tested.","tokens_in":24097,"tokens_out":11783,"duration_ms":97028,"significance":"If the reduction results are correct, the paper provides a clean geometric mechanism for nonlinear dynamics from linear evolution and a variational scheme for time-dependent problems. The restriction computations are explicit, no parameters are fitted, and the Darboux-coordinate analysis in §4.1–4.2 gives integrable finite-dimensional systems. However, the advertised generalization to non-invariant submanifolds lacks any validation, so the significance of that part is presently conditional.","major_comments":[{"comment":"The central new claim—that the pulled-back Lagrangian dynamics approximates the unitary flow when the submanifold is not invariant—is not supported. The manuscript states verbatim that \"an extensive comparison between solutions of the equations of the motion on the submanifold of Gaussian states and their quantum evolution is beyond the scope of this paper and will be addressed elsewhere,\" and the conclusion repeats that approximation procedures are future work. No error estimate and no numerical comparison are provided. Without such a check, the anharmonic-oscillator equations are just one particular dynamical system on the parameter manifold, and the abstract's claim that the Lagrangian procedure \"can be extended also to non-invariant submanifolds\" as an approximation is unsubstantiated. Please add a quantitative comparison with the exact evolution (for example, the state overlap between the reduced Gaussian state and the projected Schrödinger state) or clearly present the approximation property as a conjecture.","section":"Section 4, Eq. (4.60) and paragraphs following it"},{"comment":"The two displayed equations for the real part of the Gaussian-width parameter have opposite signs for the ω² term. Substituting a = a_R + i a_I into Eq. (2.36) gives ˙a_I = 2(a_I² - a_R²) + ω²/2, in agreement with Eq. (4.42), so Eq. (2.37) is wrong as printed. This error also makes Eq. (2.37) inconsistent with the solutions (2.38)–(2.40), which correspond to the plus sign. The sign in Eq. (2.37) should be corrected.","section":"§2.2, Eq. (2.37) vs. §4.2, Eq. (4.42)"}],"minor_comments":[{"comment":"The coefficient of db_R contains a spurious term ω b_R/(4a_R²) in addition to the correct ω² b_R/(4a_R²). Differentiating Eq. (4.58) with respect to b_R yields only the latter term; the printed differential is therefore inconsistent with the equations of motion (4.60), which contain no corresponding ω term in the evolution of b_I.","section":"§4.2, Eq. (4.59)"},{"comment":"The text says that equality of the two reduction procedures on an invariant submanifold \"is, actually, a general result which will be shown in the rest of this section,\" but the subsequent argument only proves that the pulled-back forms and energy satisfy the commutation relations (4.20). The proof that the restricted Schrödinger vector field coincides with the Hamiltonian vector field defined by (4.22) is only sketched; please state this argument explicitly.","section":"Section 4, before Example 1"},{"comment":"There is a typo in the second bullet: \"insotropic\" should read \"isotropic.\" Similar small typos (e.g., \"corrsponding\" in Example 1) should be corrected.","section":"Section 4, bullet list"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is a competent contribution to the geometric reduction literature, and the invariant-submanifold computations appear sound. The main barrier is the gap between the advertised dynamical variational method and the evidence: the only non-invariant example is not validated, and the paper itself defers that validation to future work. I would support publication after the authors either supply the missing comparison or substantially soften the approximation claim, and after the sign error in Eq. (2.37) is corrected."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a solid but uneven paper, and the central mechanism is not new — it is already in the authors' own refs [1] and [21]. The contribution is the worked examples: explicit Gaussian-state reduction for free and harmonic oscillators, Darboux coordinates, integrability statements, and a Grassmann-variable illustration. Those parts are mostly correct and honestly presented; the citation pattern is fine.\n\nWhat is genuinely good: Sections 2 and 4.1–4.2 reproduce the known nonlinear Riccati-type dynamics for Gaussian states, with Darboux coordinates and two constants of motion in involution. The supermanifold example in Section 4.3 is a nice check that the pullback procedure works beyond ordinary coherent states. The authors also flag explicitly that the non-invariant-submanifold case is only an approximation and that the comparison with exact quantum evolution is future work. That is honest, but it means the abstract's advertised generalization rests on an untested claim.\n\nSoft spots, in order of importance. First, the missing fidelity check for the anharmonic oscillator: Eq. (4.60) is derived, and then the paper says an extensive comparison between the reduced and exact dynamics is beyond scope. Without an error estimate, a numerical check, or a theorem, there is no reason to believe the pulled-back Lagrangian flow stays close to the Schrödinger flow or its projection. The paper can be read as a derivation of candidate reduced equations, but not as a validated approximation method.\n\nSecond, a sign inconsistency: Eq. (2.37) has −ω²/2 in the equation for ˙a_I, while Eq. (4.42) has +ω²/2. The plus sign is the correct one — it follows from Eq. (2.36) and is required for the Darboux equations (4.47)–(4.48) to hold. Since both sections are supposed to describe the same invariant dynamics, one of them is wrong; likely (2.37) is a typo, but it needs fixing.\n\nBottom line: this deserves a serious referee. The computations are explicit and checkable, and the paper is a useful reference for Gaussian-state reduction and the time-dependent variational principle. But in its current form the central advertised claim is unsupported. My recommendation: send to peer review; require the sign fix; and ask the authors either to add a numerical comparison for the anharmonic oscillator or to weaken the abstract and state clearly that the approximation property is conjectural.","headline":"The invariant-submanifold reduction is careful and mostly correct, but the advertised extension to non-invariant submanifolds is never checked against the exact Schrödinger flow, and the harmonic-oscillator section contains a sign slip.","tokens_in":24594,"tokens_out":6184,"would_cite":true,"duration_ms":55121,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["81R30","53D22","70H03","81Q05"],"pacs":["03.65.-w","02.40.-k"],"model":"deepseek-v4-flash","headline":"A linear quantum flow, restricted to Gaussian or coherent states, becomes generically nonlinear on the parameter manifold; a pulled-back Lagrangian handles non-invariant families.","keywords":["nonlinear dynamics","unitary evolution","coherent states","Gaussian states","squeezed states","Lagrangian formalism","symplectic reduction","variational approximation"],"falsifier":"Run the anharmonic-oscillator Schrödinger equation numerically from a normalized Gaussian initial state with $\\lambda>0$ and compare the time-dependent width, centroid, and overlap with the trajectory of the reduced system (4.60). If the reduced trajectory departs from the exact evolution on a time scale comparable to the harmonic period, the paper's advertised approximation scheme for non-invariant families is not supported.","tokens_in":23580,"feed_emoji":"⚛️","tokens_out":13360,"duration_ms":117714,"temperature":0.7,"pith_summary":"This paper establishes that linear quantum dynamics can acquire nonlinearity when it is confined to a submanifold of states. If a family of trial states—Gaussian states, squeezed and correlated states, or coherent states—is invariant under the unitary flow of the Schrödinger equation, the induced motion on the family's parameters is typically nonlinear, even though the underlying evolution is linear. When the family is not invariant, the paper shows that a Lagrangian description of the Schrödinger equation can be pulled back to the parameter manifold, producing a finite-dimensional dynamics that is only an approximation of the true flow. The payoff is a geometric mechanism for the emergence of classical-like nonlinear dynamics from linear quantum mechanics, and a dynamical generalization of the stationary variational method.","feed_headline":"Linear quantum flow becomes nonlinear on restricted state families","feed_subtitle":"A pulled-back Lagrangian gives approximate finite-dimensional dynamics when the trial family is not invariant.","key_machinery":"The central object is the pair $(\\tilde\\omega_L,\\tilde E_L)$ obtained by pulling back, through the tangent map of the immersion $i:M\\to\\mathcal{H}$, the Lagrangian two-form and the energy of the projective Schrödinger dynamics. The equation $i_{\\tilde\\Gamma_l}\\tilde\\omega_L=-\\tilde d\\tilde E_L$ then defines the induced vector field on the parameter manifold; when the pulled-back two-form is degenerate, its kernel generators (dilations and phase rotations) must be handled separately, and a parallel-transport condition fixes the normalization and phase. In the invariant case, the same result is obtained by directly restricting the original Schrödinger vector field, and the two procedures provably coincide; in this way the Lagrangian machinery supplies a dynamical generalization of the variational method for trial states.","core_discovery":"On the paper's own terms, the central claim is that a one-parameter unitary group $\\phi_t$ on a Hilbert space $\\mathcal{H}$, restricted to an immersed submanifold $i(M)\\subset\\mathcal{H}$ of states, induces a one-parameter group $\\tilde\\phi_t$ on $M$, and this induced dynamics is generically nonlinear regardless of whether the embedding $i$ itself is nonlinear. For invariant submanifolds, the restriction of the Schrödinger vector field directly gives $\\tilde\\phi_t$; for non-invariant submanifolds, the covariant Lagrangian procedure defines a vector field $\\tilde\\Gamma_l$ on $TM$ by $i_{\\tilde\\Gamma_l}\\tilde\\omega_L=-\\tilde d\\tilde E_L$, where $\\tilde\\omega_L=(Ti)^*\\omega_L$ and $\\tilde E_L=(Ti)^*E_L$, and this is an approximation of the full evolution within the chosen family. The paper verifies the invariant case on squeezed and correlated Gaussian states for the free particle and harmonic oscillator, obtaining nonlinear equations for the Gaussian parameters; it verifies the non-invariant case on the anharmonic oscillator, obtaining the finite-dimensional system (4.60). A further claim is that the nonlinearity of the induced dynamics is unrelated to the nonlinearity of the embedding: for spin-coherent states and for bosonic and fermionic oscillator coherent states, the induced motion is simply $z\\mapsto z e^{i\\omega t}$.","pith_inferences":["If the anharmonic reduction proves quantitatively accurate, the procedure offers a practical route to simulate infinite-dimensional quantum dynamics by integrating finite ODEs on a chosen trial-state family; adding more parameters should systematically improve fidelity, much as enlarging a variational ansatz improves ground-state energies.","The symplectic coordinate reformulation of the Gaussian induced dynamics, with an inverse-square potential in the variance variables, hints that the reduced systems may be completely integrable in broader settings; this could be tested by searching for additional Poisson-commuting invariants beyond the two found for the free and harmonic cases.","The same pulling-back construction could be applied to open-system dynamics: the paper sketches an invariant equatorial disc in the qubit state space for a non-Markovian qubit evolution, suggesting a route to classical-like non-Markovian dynamics on parameter manifolds."],"forward_implications":["A linear quantum flow restricted to an invariant state family is generically nonlinear: Gaussian states for the free particle and the harmonic oscillator obey the nonlinear systems (2.15) and (2.36), so nonlinearity can arise from confinement to a submanifold rather than from nonlinear fundamental laws.","When the submanifold is invariant, the direct restriction procedure and the pulled-back Lagrangian procedure give the same dynamics on the parameter manifold.","For non-invariant families, the pulled-back Lagrangian yields a finite-dimensional system of ODEs, such as equations (4.60) for the anharmonic oscillator, generalizing the variational method from stationary problems to time-dependent ones.","The induced dynamics can be linear even when the embedding is nonlinear: spin-coherent states and bosonic and fermionic oscillator coherent states all rotate as $z\\mapsto z e^{i\\omega t}$ under their natural Hamiltonians, so the nonlinearity of the embedding and of the induced flow are independent.","The same Lagrangian reduction works for anticommuting-valued parameter spaces, where it reproduces fermionic anticommutation relations through a symmetric bracket, indicating that the construction covers fermionic systems as well."],"supporting_citations":[{"why":"Establishes that restricting a dynamics to a nonlinear invariant submanifold can produce nonlinear systems, the starting idea of the paper.","marker":"[1]"},{"why":"Defines squeezed and correlated Gaussian states used as the main invariant submanifold in the free-particle and oscillator examples.","marker":"[5]"},{"why":"Introduces spin-coherent states used to exhibit a nonlinear embedding with linear induced dynamics.","marker":"[8]"},{"why":"Provides the standard harmonic-oscillator coherent states whose induced motion under the oscillator Hamiltonian is a rotation of the parameter.","marker":"[10]"},{"why":"Introduces f-oscillators and nonlinear coherent states used in the example showing that only a constant deformation preserves a linear coherent-state evolution.","marker":"[12]"},{"why":"Supplies the Lagrangian description of the time-dependent variational principle for the Schrödinger equation that the paper pulls back to the parameter manifold.","marker":"[21]"},{"why":"Documents that Gaussian-type coherent states are not invariant under anharmonic-oscillator evolution, motivating the non-invariant Lagrangian approximation.","marker":"[26]"},{"why":"Provides the geometric formalism of Lagrangian supermechanics used for the anticommuting-variable example.","marker":"[28]"}],"fun_headline_variants":["Linear quantum evolution becomes nonlinear on state submanifolds","Restricting linear Schrödinger flow induces nonlinear dynamics","Submanifold restriction turns linear unitary flow nonlinear","Induced dynamics from linear quantum flow can be nonlinear","Linear quantum dynamics becomes nonlinear on state families"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the motion computed on a non-invariant family of trial states is a faithful approximation of the actual Schrödinger evolution; the paper gives the reduced equations for the anharmonic oscillator, but explicitly leaves the detailed comparison with the full quantum evolution to future work.","fun_headline_variants_meta":{"raw":{"variants":["Linear quantum evolution becomes nonlinear on state submanifolds","Restricting linear Schrödinger flow induces nonlinear dynamics","Submanifold restriction turns linear unitary flow nonlinear","Induced dynamics from linear quantum flow can be nonlinear","Linear quantum dynamics becomes nonlinear on state families"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000655,"raw_usage":{"total_tokens":3029,"prompt_tokens":1004,"completion_tokens":2025,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":620,"completion_tokens_details":{"reasoning_tokens":1951}},"tokens_in":620,"tokens_out":2025,"duration_ms":15398,"temperature":1.0,"reasoning_tokens":1951,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T14:06:11.095869+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the anharmonic-oscillator Schrödinger equation numerically from a normalized Gaussian initial state with $\\lambda>0$ and compare the time-dependent width, centroid, and overlap with the trajectory of the reduced system (4.60). If the reduced trajectory departs from the exact evolution on a time scale comparable to the harmonic period, the paper's advertised approximation scheme for non-invariant families is not supported.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Establishes that restricting a dynamics to a nonlinear invariant submanifold can produce nonlinear systems, the starting idea of the paper."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Defines squeezed and correlated Gaussian states used as the main invariant submanifold in the free-particle and oscillator examples."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Introduces spin-coherent states used to exhibit a nonlinear embedding with linear induced dynamics."},{"cited_title":"Glauber Photon Correlations","cited_arxiv_id":null,"evidence_quote":"Provides the standard harmonic-oscillator coherent states whose induced motion under the oscillator Hamiltonian is a rotation of the parameter."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Introduces f-oscillators and nonlinear coherent states used in the example showing that only a constant deformation preserves a linear coherent-state evolution."},{"cited_title":"Kramer and M","cited_arxiv_id":null,"evidence_quote":"Supplies the Lagrangian description of the time-dependent variational principle for the Schrödinger equation that the paper pulls back to the parameter manifold."},{"cited_title":"Krivoshlykov, V.I","cited_arxiv_id":null,"evidence_quote":"Documents that Gaussian-type coherent states are not invariant under anharmonic-oscillator evolution, motivating the non-invariant Lagrangian approximation."},{"cited_title":"Ibort and J","cited_arxiv_id":null,"evidence_quote":"Provides the geometric formalism of Lagrangian supermechanics used for the anticommuting-variable example."}],"review_version":1}