{"id":"efb7d16d-deb6-4b7b-8d51-06d5c0811abb","arxiv_id":"2608.01878","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":5.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"Collective Hamiltonians of any Hamiltonian action of a compact semisimple Lie group form a superintegrable system, with action variables given by the Guillemin-Sternberg torus momentum map.","lead":"This paper proves that Hamiltonians built only from the conserved charges of a compact symmetry form a 'superintegrable' system, meaning they carry more hidden conserved quantities than ordinary integrability allows. It shows these extra quantities come from a classical torus action, which also supplies the angle variables used in quantization.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"No significant objection identified: central claim is well-supported; cited principal-face theorem is a legitimate input.","rationale":"The reader's verdict is CONDITIONAL with high confidence, and the paper's central claim is indeed a theorem about collective superintegrability. I looked specifically for a step that could break the equality ddim(H)+ddim(F)=dim(M). The principal-face theorem (Theorem 2.9) is the deepest input: it gives the dense open submanifold M_sigma, the smooth sweeping map on Sigma_sigma, and hence the Guillemin-Sternberg torus action. But this is a cited theorem, not an unstated assumption; its hypotheses match the paper's compact-G setting. The proof of ddim(H)=dim(M)-D in Lemma 3.3 is sound, relying only on the standard fact that invariant functions separate tangent directions along the torus action. The lower bound ddim(F)>=D in Lemma 3.5 is constructed correctly; the missing upper bound is elementary and does not challenge the theorem. The smooth extension by zero outside M*_sigma is legitimate because the constructed functions have compact support contained in M*_sigma. The completeness-of-flows point in Definition 1.1 follows from compactness of G. Thus no new load-bearing concern emerges. I retain the reader's CONDITIONAL rather than upgrading to ACCEPT only because the paper leaves a few short arguments implicit and defers the Thimm-chain generalization; these are revision-level matters, not threats to the central claim.","tokens_in":13071,"tokens_out":23074,"duration_ms":277945,"concrete_test":"Write out the omitted upper bound for Lemma 3.5: fix x in M*_sigma and show that any F in F satisfies dF(x) in (span{X_H(x) : H in H})^deg, whose dimension is dim(M)-ddim(H) = D by Lemma 3.3. Together with the D independent elements constructed in Lemma 3.5, this closes the equality ddim(F)=D and hence Theorem 3.2. As a second check, verify that the cited Theorem 2.9 applies verbatim to the noncompact examples in Section 1, in particular to T*Q with Q a homogeneous G-space, by confirming that the hypotheses in Lane's Theorem 2 and Hilgert–Neeb–Plank Lemmas 6.7-6.9 do not require compactness of M or properness of J.","verdict_should_be":"UNCHANGED","load_bearing_attack":"No load-bearing internal flaw found. The closest candidate is the dependence of Theorem 3.2 on Theorem 2.9 (the principal open face structure), which is imported from Lane and Hilgert–Neeb–Plank rather than proved here. If that theorem required compactness of M or properness of the momentum map—hypotheses not stated in the paper—the density of M_sigma and the functional-dimension computations would fail for advertised examples such as T*Q. However, the paper's standing assumption is a compact connected group G, which makes the G-action proper, and the quoted theorem is precisely the established convexity/stratification result for that setting. The reader's other concerns are minor and fillable: Lemma 3.5 proves only ddim(F) >= D, but the upper bound follows because any F in the centralizer satisfies dF(x) in the annihilator of the span of Hamiltonian vector fields of H, whose dimension is ddim(H) = dim(M)-D by Lemma 3.3; the smooth-by-zero extension is justified by compact support of the constructed functions inside M*_sigma; and completeness of collective Hamiltonian flows follows from compactness of G via p(t)=exp(t grad h(J(p0))) . p0. None of these rises to a load-bearing defect.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proves that for a Hamiltonian action of a connected compact Lie group G with simple Lie algebra on a symplectic manifold (M,ω), the Abelian Poisson algebra H=J^*(C^∞(g*)^G) of collective Hamiltonians and its centralizer F satisfy ddim(H)+ddim(F)=dim(M). Under a non-triviality condition, this yields a superintegrable system in the sense of Definition 1.1. The proof uses the Guillemin–Sternberg torus action associated with the principal open face σ of the momentum map: the components of the GS momentum map μ_σ=Ψ_σ∘J_σ provide the independent collective Hamiltonians on a dense open submanifold, while the quotient M*_σ/T_σ supplies the independent centralizer functions. The paper further shows that μ_σ gives action variables in generalized action-angle coordinates, and compares the result with the earlier complete-algebra approach of Bolsinov and Jovanović.","tokens_in":13156,"tokens_out":5448,"duration_ms":70493,"significance":"If the result stands, it gives a clean geometric explanation of collective superintegrability: the equality of functional dimensions is derived from the geometry of the GS torus action rather than imposed. The paper also contributes a useful new ingredient by identifying the action variables explicitly as the momentum map of the GS torus, which is nontrivial and potentially useful for quantization and reduction. The Lie-theoretic preparation in Section 2 is careful and the cited structural input (Theorem 2.9 from Lane and Hilgert–Neeb–Plank) is appropriate: the standing compactness of G makes the action proper, so the stress-test concern about the imported principal-face theorem does not land. The proof is not machine-checked, but the logical chain is transparent and the missing steps identified below are local and fixable.","major_comments":[{"comment":"The proof of Lemma 3.5 establishes only the lower bound ddim(F) ≥ D by constructing D independent centralizer functions. The upper bound is not stated. It is needed for the equality ddim(F)=D and can be supplied by a short argument: for any F∈F and x∈M*_σ, dF(x) must annihilate the span of the Hamiltonian vector fields of H, whose dimension is dim(M)-D by Lemma 3.3 and Corollary 3.4; hence ddim(F) ≤ D. In addition, the smooth extension by zero outside M*_σ needs a brief justification: since M*_σ/T_σ is open in M/T_σ, the functions f_i can be chosen with compact support in the quotient coordinate chart, so their preimages are closed in M; the Poisson bracket {F_i,H} then vanishes on the dense open M*_σ and, being continuous, vanishes identically.","section":"§3.1, Lemma 3.5"},{"comment":"The proof that the non-triviality condition (3.1) implies the strict inequality ddim(H)<dim(M)/2 in (1.1) is too telegraphic. The statement 'ddim(H)≤rank(G) and 2 rank(G) is smaller than the minimal dimension ... except for minimal coadjoint orbits of SU(n)' needs to be formulated precisely: one should show that (3.1) excludes not only coadjoint orbits of G but also any Hamiltonian G-space whose principal T_σ orbits would be trivial, and then argue that dim(M)>2rank(G) in the remaining cases. As written, this is a plausible sketch but not a complete derivation; since (1.1) is part of the definition of superintegrability, the gap should be filled.","section":"§3.1, Eq. (3.1)–(3.2)"}],"minor_comments":[{"comment":"In Eq. (3.12), the last sum uses 'dp_k ∧ dq_k' with the summation index k, which clashes with the fixed integer k=1/2(dim(M)-2ℓ). Rename either the index or the constant.","section":"§3.2, Theorem 3.7"},{"comment":"The inclusion H⊂A⊂F is correct, but it may be worth a sentence explaining why J^*(C^∞(g*)^G) is contained in the centralizer of C^∞(M)^G; this is not obvious to all readers and follows from the G-equivariance of the momentum map.","section":"§3.3, Eq. (3.18)"},{"comment":"In the proof of Theorem 2.13, the statement that μ_σ is a Poisson map because it is a composition of two Poisson maps is correct, but the second factor Ψ_σ:Σ_σ→t_σ is not literally a Poisson map unless t_σ is identified with its dual; the authors may want to spell out the identification to avoid a minor confusion.","section":"§2.3, Theorem 2.13"},{"comment":"The phrase 'ddim(H) + ddim(F) = dim(M)' is stated as the crucial equality; it may help to note explicitly that Definition 1.1 requires both ddim(H) and ddim(F) to be well-defined on a common dense set, which is established later in the proofs of Lemmas 3.3 and 3.5.","section":"§1, Introduction"}],"recommendation":"minor_revision","confidential_remarks":"The paper is within the scope of the journal and builds on legitimate cited results. The main theorem is convincing; the missing upper bound in Lemma 3.5 and the terse justification of (3.2) are local proof gaps that can be repaired by the authors without changing the framework. No concerns about novelty or citation practice arose."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The main thing to know: this is a new proof of a known phenomenon. Bolsinov and Jovanović already proved that collective Hamiltonians are superintegrable in a broader framework, and the paper says so explicitly, calling the compact-group results \"essentially equivalent.\" What's genuinely new is the route: the Guillemin–Sternberg torus action gives a clean computation of ddim(H)+ddim(F)=dim(M), and Theorem 3.7 turns the GS momentum map into explicit action variables. That last part is the real payoff and it's a solid contribution to the superintegrability literature.\n\nThe proof is sound in substance. Lemma 3.3 correctly computes ddim(H)=dim(M)-D using the rank of the GS momentum map on the dense open M*_σ. Lemma 3.5 constructs D independent centralizer elements. Proposition 3.6 connects the rank to isotropy groups. The paper leans on the principal-open-face theorem from Lane and Hilgert–Neeb–Plank; that's an imported structural input, but a legitimate one, not a hidden assumption.\n\nThe reader's four concerns are all real, and all fixable. First, Lemma 3.5 proves only the lower bound ddim(F)≥D; the equality needs an unstated annihilator argument (the differentials of centralizer functions lie in the annihilator of the Hamiltonian vector fields of H, whose dimension is dim(M)-D). Second, the smooth-by-zero extension of the F_i outside M*_σ needs a compact-support justification. Third, completeness of the Hamiltonian flows follows from compactness of G via p(t)=exp(t∇h(J(p0)))·p0, but that's never stated. Fourth, the Thimm-chain generalization is deferred to a future paper. None of these threaten the central theorem, but they should be cleaned up.\n\nI also noticed the argument around the non-triviality condition (3.1) races past the SU(n) minimal orbit exception; it's fine, just a bit clipped.\n\nI'd send this to a referee. It's an honest, well-written paper that adds a clear proof strategy and explicit action variables to a known result. The revision is straightforward. In a reading group on integrable systems or moment maps, it would be a useful session.","headline":"A competent new proof that collective Hamiltonians are superintegrable via the Guillemin–Sternberg torus action, honestly building on Bolsinov–Jovanović, with a few fillable gaps.","tokens_in":13871,"tokens_out":2783,"would_cite":true,"duration_ms":28795,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["70H06","53D20","37J35"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper proves that for every Hamiltonian action of a connected compact Lie group with simple Lie algebra, the collective Hamiltonians form a superintegrable system, and the Guillemin–Sternberg torus action supplies the action variables.","keywords":["collective superintegrability","Guillemin–Sternberg torus action","Hamiltonian group action","momentum map","action-angle coordinates","symplectic manifold","coadjoint invariants","functional dimension"],"falsifier":"Exhibit a Hamiltonian action of a connected compact simple Lie group on a compact symplectic manifold satisfying the non-triviality condition for which direct computation gives ddim(H) + ddim(F) > dim(M), or for which the GS torus momentum map fails to be smooth on a dense open subset. The paper predicts neither can happen; for instance, one could compute the explicit functional dimensions for the cotangent bundle of a sphere with the standard orthogonal-group action and check the equality.","tokens_in":12736,"feed_emoji":"🎯","tokens_out":4849,"duration_ms":52405,"temperature":0.7,"pith_summary":"The paper seeks to establish that collective Hamiltonians—functions pulled back from coadjoint invariants through the momentum map of a Hamiltonian group action—are always superintegrable, not merely integrable. It proves that the Abelian Poisson algebra of these collective Hamiltonians together with its centralizer satisfies the functional-dimension equality that defines superintegrability. The proof runs through the Guillemin–Sternberg torus action, a Hamiltonian torus action on a dense open submanifold that commutes with the original group action and whose momentum map components are exactly the collective Hamiltonians' differentials. This torus action also makes the action variables explicitly computable, which is usually the hardest part of a superintegrable system. If correct, the result gives a uniform explanation of collective superintegrability and opens a direct route to action variables for reduced systems such as spin many-body models.","feed_headline":"Every compact symmetry makes collective Hamiltonians superintegrable","feed_subtitle":"The Guillemin–Sternberg torus action provides the action variables that prove the equality.","key_machinery":"The key object is the Guillemin–Sternberg (GS) torus action. On a dense open G-invariant submanifold M_σ, the sweeping map Ψσ—which sends each coadjoint orbit to its representative in the principal open face σ of the fundamental Weyl chamber—composed with the momentum map J defines a momentum map μσ for a Hamiltonian T_σ action. The principal open face is the unique face of the Weyl chamber whose G-orbit stratum contains a dense open part of the momentum image. Smoothness of the sweeping map on that stratum is what makes μσ smooth, and the components of μσ generate the collective Hamiltonians' differentials on a dense open set. The dimension count ddim(H) + ddim(F) = dim(M) is then a direct","core_discovery":"The central claim is Theorem 3.2: for a Hamiltonian action of a connected compact Lie group G with simple Lie algebra on a connected symplectic manifold (M, ω), the Abelian Poisson algebra H generated by J*(C∞(g*)^G) and its centralizer F satisfy ddim(H) + ddim(F) = dim(M). Under a non-triviality condition excluding coadjoint orbits, this constitutes a superintegrable system in the sense of Definition 1.1. Theorem 3.7 then shows that the momentum map of the Guillemin–Sternberg torus action, after passing to a factor torus acting freely, provides generalized action-angle coordinates: the action variables are components of this momentum map, while the transversal coordinates come from the cent","pith_inferences":["Editorial inference: the same GS-torus machinery should provide action variables for Thimm-style chains of subgroups without additional work, making the superintegrable-versus-Liouville dichotomy for such chains effective rather than merely existential.","Editorial inference: since the proof only needs the sweeping map to be smooth on the principal stratum, the functional-dimension equality might hold under weaker smoothness hypotheses; a natural test is whether continuous or piecewise-smooth extensions of the sweeping map still yield the dimension count.","Editorial inference: the explicit action variables suggest a direct route to semiclassical quantization, with Bohr–Sommerfeld conditions for collective superintegrable systems read off from the GS momentum map.","Editorial inference: the reliance on the principal open face structure implies that, outside the compact setting or when that face degenerates, collective superintegrability could fail; this is a testable boundary of the theorem."],"forward_implications":["Every Hamiltonian G-manifold with simple compact G, excluding coadjoint orbits, carries a superintegrable system of collective Hamiltonians; the rank equals the difference between the typical isotropy dimensions for the action on M and on the momentum image.","Action variables for these systems are given explicitly by the momentum map of the GS torus action; they are continuous on the whole phase space and smooth on a dense open submanifold.","The proof extends to reductive compact groups with minor modifications, and to chains of subgroup actions, where the combined invariant Hamiltonians are again either Liouville or superintegrable.","Because the action variables come from a momentum map, they survive Hamiltonian reduction, so reduced many-body systems inherit explicit action variables."],"fun_headline_variants":["GS torus action supplies action variables for superintegrability","Compact symmetry proves collective superintegrability","Torus action gives action variables for superintegrable systems","Guillemin-Sternberg torus yields superintegrable collective systems","GS torus action proves superintegrability of collective Hamiltonians"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The argument depends on a structural fact, quoted from the literature, that the momentum image has a unique principal open face of the Weyl chamber whose stratum is dense open and on which the sweeping map is smooth; if that fact failed, the dimension count would not go through.","fun_headline_variants_meta":{"raw":{"variants":["GS torus action supplies action variables for superintegrability","Compact symmetry proves collective superintegrability","Torus action gives action variables for superintegrable systems","Guillemin-Sternberg torus yields superintegrable collective systems","GS torus action proves superintegrability of collective Hamiltonians"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000936,"raw_usage":{"total_tokens":3852,"prompt_tokens":770,"completion_tokens":3082,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":514,"completion_tokens_details":{"reasoning_tokens":2999}},"tokens_in":514,"tokens_out":3082,"duration_ms":22938,"temperature":1.0,"reasoning_tokens":2999,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-04T19:06:21.959821+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Exhibit a Hamiltonian action of a connected compact simple Lie group on a compact symplectic manifold satisfying the non-triviality condition for which direct computation gives ddim(H) + ddim(F) > dim(M), or for which the GS torus momentum map fails to be smooth on a dense open subset. The paper predicts neither can happen; for instance, one could compute the explicit functional dimensions for the cotangent bundle of a sphere with the standard orthogonal-group action and check the equality.","supporting_citations":[],"review_version":1}