{"id":"6a4f248f-6435-4c82-89e3-126ba34c5061","arxiv_id":"2411.17988","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A single theorem about morphisms of Manin pairs unifies the integration of Poisson manifolds, Dirac structures, and Lie bialgebroids to groupoids, and extends them to non-simply-connected settings.","lead":"Mathematicians found a single general construction that covers several known ways of integrating geometric structures like Poisson manifolds and Dirac structures into groupoids. The new framework makes the results simpler and also works in cases where earlier proofs did not.","discovery_kind":"unification","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Appendix A equation (27) is sign-inconsistent with α from (19); the bracket-closure proof of Theorem 3.10 as written is incomplete.","rationale":"The reader's identified weakest assumption, the nontriviality of \\hat G-equivariance for non-source-simply-connected groupoids, is a legitimate scope condition but not an internal flaw: the theorem explicitly assumes it, and the paper explains how this assumption behaves in examples. The more load-bearing concern is an internal sign inconsistency between the definition of α (eq. 19) and the explicit form of the extension φ(ζ) (eq. 27) used in the Courant bracket calculations of Appendix A. Because those calculations are exactly what establishes that R is a Dirac structure, the sign error affects the central theorem's proof as written. The test proposed would settle whether the proof can be repaired by a sign correction; without such a correction, the theorem's proof has a gap. This leads to the same recommendation as the reader (CONDITIONAL), but for a different reason.","tokens_in":28709,"tokens_out":63373,"duration_ms":553155,"concrete_test":"Recompute the final bracket-closure calculation in Appendix A after changing (27) to φ(ζ) = ((ζ,ζ)+O(y), a_E(ζ) - Σ_i⟨ζ,ξ_i⟩dy_i + O(y)). Verify on gr0(t,s) that for each frame element ξ_i, ⟨[[φ(ζ),φ(ζ')]], ψ(ξ_i)⟩ equals ⟨φ([[ζ,ζ']]), ψ(ξ_i)⟩. If the derivative terms still fail to cancel, the Dirac-structure proof for R is incomplete and Theorem 3.10 lacks a rigorous proof.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"In Theorem 3.10, the units of R are the graph of α = (a_E, -pr_A*) (eq. 19). But Appendix A defines the extension φ(ζ) of φ0(ζ) = ((ζ,ζ), α(ζ)) and then writes in (27) the 1-form part of its TG component as +Σ_i ⟨ξ_i,ζ⟩ dy_i at y=0. Since pr_A*ζ = Σ_i ⟨ξ_i,ζ⟩ dy_i, this gives +pr_A*ζ, not the -pr_A*ζ required by α. Thus φ(ζ) does not restrict to a section of R on gr0(t,s) with the stated formula. More importantly, the subsequent proof that [[φ(ζ),φ(ζ')]] - φ([[ζ,ζ']]) pairs to zero with ψ(ξ_i) uses derivative terms L_{a(φ(ζ))}⟨φ(ζ'),ψ(ξ_i)⟩. With the printed sign, the two derivative terms do not cancel on gr0; the paper's claim that they vanish because a(φ(ζ)) is tangent to gr0 is insufficient, since the functions ⟨φ(ζ'),ψ(ξ_i)⟩ are not constant on gr0. Flipping the sign in (27) would make these pairings vanish on gr0 and restore the argument, so the concern is a concrete, fixable sign error rather than a conceptual failure. However, as printed, the proof of Step 8 of Theorem 3.10 is not valid.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a general integration theorem for Manin pairs. Given a Manin pair (E,A) over M, a Lie groupoid G integrating A, and a jet-groupoid action of \\hat G=J^1(G) on E satisfying natural compatibility conditions, the authors construct a canonical multiplicative Courant morphism R:(TG,TG) to (Pair(E),Pair(A)) with base map (t,s):G→M×M. The morphism is given explicitly in equation (20). The paper then derives known integration results for Poisson manifolds, twisted Dirac structures, and (quasi-)Lie bialgebroids as special cases, discusses the non-source-simply-connected situation, and develops Hamiltonian spaces and Morita equivalences in this formalism.","tokens_in":29002,"tokens_out":17820,"duration_ms":169455,"significance":"If correct, Theorem 3.10 provides a clean and genuinely unifying framework: it recovers the Mackenzie–Xu, Bursztyn–Crainic–Weinstein–Zhu, and Iglesias-Ponte–Laurent-Gengoux–Xu results from one explicit formula, and it gives a precise statement of the extra equivariance hypothesis needed when G is not source-simply connected. The paper's strengths are its explicit formula (20), the detailed appendix with the bracket computations, and the applications to quasi-symplectic groupoids and Hamiltonian spaces. However, the central proof as printed contains a sign inconsistency in Appendix A that affects the bracket-closure argument, so the main theorem is not fully established in the present version.","major_comments":[{"comment":"There is a sign inconsistency between the definition of α in (19) and the extension φ(ζ) in (27). In (19) the units relation is α(ζ)=(a_E(ζ), -pr_{A^*}ζ), but immediately before (27) the text states pr_{A^*}ζ = Σ_i ⟨ξ_i,ζ⟩dy_i, and then (27) gives the TG-component of φ(ζ) as a_E(ζ)+Σ_i ⟨ξ_i,ζ⟩dy_i, i.e. with a plus sign. Thus φ(ζ) does not restrict to a section of R on gr0(t,s). Consequently, the subsequent computation of [[φ(ζ),φ(ζ')]] does not establish that the bracket lies in R. In the pairing identity for ⟨[[φ(ζ),φ(ζ')]],ψ(ξ_i)⟩, the two derivative terms L_{a(φ(ζ))}⟨φ(ζ'),ψ(ξ_i)⟩ and L_{a(φ(ζ'))}⟨φ(ζ),ψ(ξ_i)⟩ cancel only with the opposite sign in (27); the observation that a(φ(ζ)) is tangent to gr0(t,s) is insufficient because the paired functions are not constant on gr0(t,s). Replacing the '+' in (27) with '−' appears to repair the computation, so this is likely a fixable sign error, but as printed the proof of Step 8 of Theorem 3.10 is incomplete.","section":"Appendix A, Eq. (27) and Eq. (19)"}],"minor_comments":[{"comment":"The infinitesimally multiplicative version R0 is stated as a theorem, but the proof is omitted with only a sentence saying it is similar to Theorem 3.10. If R0 is intended as a new result, it needs a proof or a precise reference; if it is only a recap of the first author's thesis, that should be stated explicitly.","section":"Section 3.5"},{"comment":"The proof writes α=(a_E,pr_{A^*}) without the minus sign appearing in (19). The surjectivity conclusion is unaffected, but the notation should be harmonized with the main definition of α.","section":"Section 4.3, proof of Proposition 4.3"},{"comment":"The text reads 'quasi-Lie bialgeboid'; this should be 'quasi-Lie bialgebroid'. Similar typographical inconsistencies between 'bialgebroid' and 'bialgeboid' appear elsewhere.","section":"Section 4.4, first paragraph"},{"comment":"The notation O(y) is used for sections of two different bundles without specifying that it means terms vanishing at y=0 in the appropriate fiber bundle; a brief clarification would improve readability.","section":"Appendix A, proof of Lemma A.2"}],"recommendation":"major_revision","confidential_remarks":"The sign error in Appendix A is the main technical obstacle. It seems eminently fixable, and the conceptual framework of the paper is attractive and likely correct. I would also ask the authors to clarify the status of R0 in Section 3.5 during revision. The paper is within scope for a differential-geometry journal and should be reconsidered after the sign issue is resolved."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a good paper that deserves a serious referee, but the proof as printed has a sign error in Appendix A that breaks the main bracket-closure argument. The error is local and fixable — flip a sign in equation (27) and the argument goes through — but as written, Step 8 of Theorem 3.10 is not valid.\n\nWhat's genuinely new: the explicit formula (20) for the integration morphism, the treatment of non-source-simply-connected groupoids via the \\hat G-action, and the Hamiltonian-space interpretation in Section 5. The paper correctly recovers known integration results (Mackenzie–Xu, BCWZ, quasi-Lie bialgebroids) as special cases, and the proof structure is transparent. The discussion of fat groupoids and the K-action in Appendix B is useful and careful.\n\nThe soft spots, in order of severity. First, the sign issue. In Appendix A, \\phi_0(\\zeta) = ((\\zeta,\\zeta), \\alpha(\\zeta)) with \\alpha = (a_E, -\\mathrm{pr}_{A^*}). But equation (27) writes the TG-component as a_E(\\zeta) + \\sum_i \\langle\\xi_i,\\zeta\\rangle dy_i. Since \\sum_i \\langle\\xi_i,\\zeta\\rangle dy_i = \\mathrm{pr}_{A^*}\\zeta, this is +\\mathrm{pr}_{A^*}\\zeta, not -\\mathrm{pr}_{A^*}\\zeta. So the extended section \\phi(\\zeta) does not restrict to a section of R over gr0(t,s). The subsequent pairing argument then fails: the derivative terms L_{a(\\phi(\\zeta))}\\langle\\phi(\\zeta'),\\psi(\\xi_i)\\rangle do not cancel with the printed sign. I checked: flipping the sign in (27) makes the restriction correct and the pairings vanish. So this is a concrete, fixable error, not a conceptual one. But a referee should not let it slide.\n\nSecond, Section 3.5 states the infinitesimal version R0 and says \"we omit the proof.\" Since the paper's whole point is to give a geometric proof of the integration theorem, leaving the infinitesimal analogue unproved is a gap. It's a minor gap if the authors can point to the thesis, but they should either include a proof or state it as a known result.\n\nThird, the \\hat G-equivariance assumption is genuinely restrictive when G is not source-simply connected. Example 3.7 is honest about this, but readers should know that the advertised extension to non-simply-connected groupoids only works when this extra structure exists. That's a limitation, not a flaw.\n\nWho this is for: Poisson geometers and Lie groupoid people. It's a useful organizing paper. My recommendation: send it to peer review, but insist the authors fix the sign error and either prove or properly cite R0. The main theorem is very likely correct, but the printed proof is not.","headline":"A useful unifying framework with a real but fixable sign error in the main proof; deserves refereeing but not acceptance as-is.","tokens_in":93,"tokens_out":5407,"would_cite":false,"duration_ms":169545,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["53D17","22A22","53D20"],"pacs":[],"model":"deepseek-v4-flash","headline":"Given a $\\hat G$-equivariant Manin pair $(E,A)$ with $G$ integrating $A$, there is a unique $\\hat G\\times\\hat G$-equivariant morphism of multiplicative Manin pairs $R:(\\mathbb{T}G,\\mathbb{T}G)\\to(\\mathrm{Pair}(E),\\mathrm{Pair}(A))$; this…","keywords":["Manin pairs","Courant algebroids","Lie groupoids","integration","Poisson groupoids","quasi-symplectic groupoids","Dirac structures","Lie bialgebroids"],"falsifier":"Take a $\\hat G$-equivariant Manin pair and compute the Dorfman bracket of two sections of the relation $R$ defined by equation (20) over $\\mathrm{gr}(t,s)$; if any bracket component leaves $R$, Theorem 3.10 is false. A concrete place to look is a non-source-simply-connected integration such as $G=T^*M/\\Lambda$ with a non-Lagrangian period subgroupoid $\\Lambda$, where Example 3.7 predicts the required $\\hat G$-action does not exist, so the theorem's hypothesis fails exactly when the Lagrangian condition is violated.","tokens_in":28508,"feed_emoji":"🔄","tokens_out":9126,"duration_ms":77743,"temperature":0.7,"pith_summary":"The paper's central claim is that a whole family of integration results in Poisson geometry are instances of one construction: given a Manin pair $(E,A)$ — a Courant algebroid $E$ with a Dirac subbundle $A$ — and a Lie groupoid $G\\Rightarrow M$ integrating the Lie algebroid $A$, the pair integrates to a distinguished morphism $R$ from the standard Courant algebroid $\\mathbb{T}G=TG\\oplus T^*G$ to the pair Courant algebroid $E\\times E$, provided a jet-groupoid action exists. When $A$ is the cotangent algebroid of a Poisson manifold, this morphism packages the multiplicative symplectic form; when $E$ is exact, it packages the multiplicative 2-form of a quasi-symplectic groupoid; when a complementary Lagrangian $B$ is chosen, it packages the multiplicative Poisson or quasi-Poisson structure. The authors prove this by writing $R$ down explicitly in terms of two maps $\\alpha$ and $\\beta$, then verifying multiplicativity through Courant-bracket computations. A sympathetic reader should care because it replaces a list of separate classical theorems with a single mechanism and clarifies what happens when the integrating groupoid is not source-simply connected.","feed_headline":"One equivariant map unifies Poisson integration theorems","feed_subtitle":"The same construction yields symplectic, quasi-symplectic, and Poisson groupoids, even for non-simply-connected G.","key_machinery":"The load-bearing object is the fat groupoid $\\hat G=J^1(G)$ of the tangent VB-groupoid $\\mathbb{T}G\\Rightarrow TM$: its elements are complements to the kernels of source and target, it acts on $A$, and by assumption it acts on the whole Courant algebroid $E$. The argument runs through two dual bundle maps, $\\alpha$ from units to units and $\\beta$ from cores to cores, whose graphs span $R|_M$; $\\hat G\\times\\hat G$-equivariance then propagates $R$ over all of $G$, while the Courant-bracket computations in Appendix A show the section span is closed under the Dorfman bracket. The jet groupoid is what carries the infinitesimal representation $\\nabla_{j^1(\\xi)}\\zeta=[[\\xi,\\zeta]]$ into a global action, and the assumption that this action exists is exactly what replaces source-simple connectivity.","core_discovery":"Let $(E,A)$ be a $\\hat G$-equivariant Manin pair over $M$, where $\\hat G=J^1(G)$ is the jet groupoid of a Lie groupoid $G\\Rightarrow M$ integrating $A$. Theorem 3.10 asserts that the map $\\alpha=(a_E,-\\mathrm{pr}_{A^*}):E\\to TM\\oplus A^*$ extends uniquely to a $\\hat G\\times\\hat G$-equivariant morphism of multiplicative Manin pairs $R:(\\mathbb{T}G,\\mathbb{T}G)\\to(\\mathrm{Pair}(E),\\mathrm{Pair}(A))$ with base map $(t,s):G\\to M\\times M$. Explicitly, $x\\sim R(\\zeta',\\zeta)$ if and only if $t_{\\mathbb{T}G}(x)=\\alpha(\\zeta')$ and $\\beta([l_{\\hat g^{-1}}(x)])=\\zeta-\\mathrm{Ad}_{\\hat g^{-1}}\\zeta'$, where $\\beta=(\\iota_A,-a_E^*):A\\oplus T^*M\\to E$ and $l_{\\hat g^{-1}}$ is left translation by any lift. The paper verifies that this relation is a Lagrangian subbundle, a subgroupoid, and closed under the Courant bracket, hence a morphism of CA-groupoids, and that uniqueness follows because any such morphism with prescribed units is determined by its core. This single theorem specializes to integrations of Poisson manifolds, twisted Dirac structures, and (quasi-)Lie bialgebroids, and the formula remains valid for integrations that are not source-simply connected.","pith_inferences":["Beyond the paper, the explicit formula (20) gives a practical integrability test: to integrate a Manin pair, one only needs to construct the $\\hat G$-action on $E$, and the relation $R$ is then explicit; this shifts the difficulty from solving for multiplicative forms to checking equivariance.","Beyond the paper, the non-source-simply-connected case suggests that obstructions to integration live in the period groupoid of $A$: Example 3.7 identifies the obstruction with a Lagrangian condition on $\\Lambda$, so one can probe integrability by asking whether the canonical representation of $J^1(A)$ integrates through the period group.","Beyond the paper, the Hamiltonian-space formalism of Section 5 suggests a reduction procedure for $\\hat G$-equivariant Manin pairs: quotients of Hamiltonian $G$-spaces should remain Hamiltonian spaces for the reduced Manin pair, paralleling quasi-Hamiltonian reduction; the paper does not develop this reduction explicitly."],"forward_implications":["Given a $\\hat G$-equivariant Manin pair, the morphism $R$ is the integration of the Manin pair: any complementary Dirac or Lagrangian subbundle $B\\subset E$ pulls back under $R$ to a multiplicative (quasi-)Poisson structure on $G$ whose target-source map is a (quasi-)Poisson map.","If $E$ is an exact Courant algebroid, $R$ is automatically exact and is encoded by a multiplicative 2-form $\\omega_G$ satisfying $d\\omega_G=s^*\\eta-t^*\\eta$, recovering the integration of $\\eta$-twisted Dirac structures to quasi-symplectic groupoids and, when $\\eta=0$ and $A\\cap TM=0$, symplectic groupoids integrating Poisson manifolds.","Choosing a Lagrangian complement $B\\subset E$ to $A$, the preimage $R^{-1}(\\mathrm{Pair}(B))$ is the graph of a multiplicative bivector field on $G$; if $B$ is a Dirac structure the bivector is Poisson, giving the classical integration result for Lie bialgebroids without requiring $G$ to be source-simply connected.","A Hamiltonian $G$-space for a $\\hat G$-equivariant Manin pair, defined as a $\\hat G$-equivariant morphism $L:(\\mathbb{T}P,\\mathbb{T}P)\\to(E,A)$, is automatically a module over $R$: whenever $x\\sim R(\\zeta',\\zeta)$ and $y\\sim L\\zeta$ with compatible sources, the action satisfies $x\\cdot y\\sim L\\zeta'$, which reproduces known Hamiltonian-space notions for quasi-symplectic groupoids and (quasi-)Poiss"],"supporting_citations":[{"why":"Supplies the Lie bialgebroid integration theorem that Theorem 3.10 generalizes and recovers.","marker":"[27]"},{"why":"Provides the twisted Dirac integration result recovered in the exact Courant algebroid case.","marker":"[5]"},{"why":"Provides the quasi-Poisson groupoid integrations recovered from quasi-Lie bialgebroids.","marker":"[15]"},{"why":"Introduces Manin triples for Lie bialgebroids and the Courant algebroid construction underlying the splitting argument.","marker":"[24]"},{"why":"Supplies the CA-groupoid formalism and the definition of Hamiltonian spaces used throughout.","marker":"[21]"},{"why":"Sets out conventions for Courant algebroids and Dirac structures, including groupoid actions.","marker":"[22]"},{"why":"Defines Courant morphisms and morphisms of Manin pairs, the language in which $R$ is a morphism.","marker":"[6]"},{"why":"Provides the fat groupoid and VB-groupoid machinery used to define $\\hat G$ and its actions.","marker":"[14]"}],"fun_headline_variants":["One theorem integrates Poisson, Dirac, and Lie bialgebroids","Manin pair morphism unifies all integration theorems","Jet groupoid yields universal Poisson integration","Single construction covers non-simply connected cases","Equivariant map integrates all Poisson-like structures"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The theorem's load-bearing premise is that the jet groupoid $\\hat G=J^1(G)$ really acts on the whole Courant algebroid $E$ as Definition 3.4 demands; for source-simply connected $G$ this is automatic, but for other groupoids it is a nontrivial condition that can fail, and without it the relation $R$ has no equivariance to get off the units.","fun_headline_variants_meta":{"raw":{"variants":["One theorem integrates Poisson, Dirac, and Lie bialgebroids","Manin pair morphism unifies all integration theorems","Jet groupoid yields universal Poisson integration","Single construction covers non-simply connected cases","Equivariant map integrates all Poisson-like structures"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000576,"raw_usage":{"total_tokens":3081,"prompt_tokens":1045,"completion_tokens":2036,"prompt_tokens_details":{"cached_tokens":1024},"prompt_cache_hit_tokens":1024,"prompt_cache_miss_tokens":21,"completion_tokens_details":{"reasoning_tokens":1961}},"tokens_in":21,"tokens_out":2036,"duration_ms":290666,"temperature":1.0,"reasoning_tokens":1961,"cache_read_input_tokens":1024,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T11:37:57.003606+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a $\\hat G$-equivariant Manin pair and compute the Dorfman bracket of two sections of the relation $R$ defined by equation (20) over $\\mathrm{gr}(t,s)$; if any bracket component leaves $R$, Theorem 3.10 is false. A concrete place to look is a non-source-simply-connected integration such as $G=T^*M/\\Lambda$ with a non-Lagrangian period subgroupoid $\\Lambda$, where Example 3.7 predicts the required $\\hat G$-action does not exist, so the theorem's hypothesis fails exactly when the Lagrangian condition is violated.","supporting_citations":[{"cited_title":"3, 445–467","cited_arxiv_id":null,"evidence_quote":"Supplies the Lie bialgebroid integration theorem that Theorem 3.10 generalizes and recovers."},{"cited_title":"Bursztyn, M","cited_arxiv_id":null,"evidence_quote":"Provides the twisted Dirac integration result recovered in the exact Courant algebroid case."},{"cited_title":"Iglesias-Ponte, C","cited_arxiv_id":null,"evidence_quote":"Provides the quasi-Poisson groupoid integrations recovered from quasi-Lie bialgebroids."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Introduces Manin triples for Lie bialgebroids and the Courant algebroid construction underlying the splitting argument."},{"cited_title":"Li-Bland and E","cited_arxiv_id":null,"evidence_quote":"Supplies the CA-groupoid formalism and the definition of Hamiltonian spaces used throughout."},{"cited_title":"5, 779–816","cited_arxiv_id":null,"evidence_quote":"Sets out conventions for Courant algebroids and Dirac structures, including groupoid actions."},{"cited_title":"Bursztyn, D","cited_arxiv_id":null,"evidence_quote":"Defines Courant morphisms and morphisms of Manin pairs, the language in which $R$ is a morphism."},{"cited_title":"Gracia-Saz and R","cited_arxiv_id":null,"evidence_quote":"Provides the fat groupoid and VB-groupoid machinery used to define $\\hat G$ and its actions."}],"review_version":1}