REVIEW 3 major objections 3 minor 28 references
On the geometry of Lagrangian one-forms
T0 review · 3 major / 3 minor · reviewed 2026-08-11 · deepseek-v4-flash
Pith's one-line read A single variational principle now derives both the equations of motion and integrability's closure relation.
desk verdict Clean phase-space reformulation of Lagrangian multiforms with a genuine variational derivation of Hamiltonian group actions; the 'any finite-dimensional system' claim outruns the proved Legendre-transform assumptions. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The engine is the exterior derivative of the phase-space one-form and the kernel distribution it defines. In coordinates the Pfaffian system is $\theta_\mu=dp_\mu+(\partial H_i/\partial q^\mu)dt^i$, $\varphi_\mu=dq^\mu-(\partial H_i/\partial p_\mu)dt^i$; the rank of the kernel is $n$ precisely when $\{H_i,H_j\}=0$, and since $dL$ is closed the distribution is Frobenius-integrable. In the Lie-group generalisation the same role is played by the Maurer-Cartan form $g^{-1}dg$, whose structure constants convert the condition into the moment-map relation.
What would settle it
Choose $H_1=p^2/2+q^2/2$ and $H_2=p$ on a two-dimensional phase space. Their Poisson bracket is nonzero, so the paper's theorem predicts that the overdetermined system (2.26)\u2013(2.28) has no solution surface; solving it directly for a curve of the form $(p(t^1,t^2),q(t^1,t^2))$ should give an inconsistency. If a local solution exists for these non-commuting Hamiltonians, the claimed equivalence between the univariational principle and Poisson involutivity would be false.
Extended reading notes
Core claim
The discovery is that the variational content of an integrable hierarchy is carried by a single phase-space one-form $L=p_\mu dq^\mu-H_i dt^i$ on $M\times \mathbb{R}^n$, rather than by a position-space Lagrangian one-form plus a separate closure condition. The univariational principle demands that every curve in an $n$-dimensional hypersurface $\Sigma\subset M\times \mathbb{R}^n$ be stationary for $S[\gamma]=\int \gamma^*L$; this is equivalent to $\gamma'\lrcorner dL=0$ for all tangent directions, which in graph coordinates becomes equations (2.26)\u2013(2.28). Substituting the flow equations into the third relation gives $\{H_j,H_i\}=0$, and conversely, the Frobenius theorem applied to the kernel of $dL$ gives local solution surfaces exactly when the $H_i$ Poisson-commute; on such surfaces the one-form is closed. In the nonabelian case, $L=\alpha-(H,g^{-1}dg)$ on $M\times G$ and the Maurer-Cartan equation turn the analogous computation into $\{H_i,H_j\}=c^k_{ij}H_k$, so $H$ is a moment map and the solutions are orbits of a Hamiltonian $G$-action.
Load-bearing premise
The claim that any finite-dimensional integrable system fits the framework rests on the inverse Legendre transform being solvable: equation (3.7) must determine the momenta as functions of positions and velocities, which the paper ensures by assuming some direction in which $\alpha^i H_i$ is convex in momenta, and the velocity-linear degenerate case is explicitly left out.
Editorial extensions
If this is right
- Every Liouville integrable hierarchy whose Hamiltonians satisfy the stated convexity condition admits a position-space Lagrangian one-form, so the previously ad hoc construction becomes systematic.
- The multi-time Euler-Lagrange equations and the closure relation are obtained from one variational principle, with no separate variation of the submanifold.
- For noncommuting flows on a Lie group $G$, the Euler-Lagrange equations of $L=\alpha-(H,g^{-1}dg)$ are $\{H_i,H_j\}=c^k_{ij}H_k$, so $H$ is a moment map and the dynamics is a Hamiltonian $G$-action.
- In local coordinates on $G$, the same equations become compatible non-autonomous Hamiltonian flows, which accommodate explicitly time-dependent conserved quantities such as $C=J-tH_0$ in the worked example.
- The construction reproduces known position-space multiforms, such as the periodic Toda-chain coefficients recovered at a special parameter value.
Reading between the lines
- If the univariational principle is taken as basic, integrability becomes the statement that $dL$ vanishes on the solution surface; a natural next step, not taken in the paper, is to read the closure relation as a flatness condition and to search for a cohomological classification of Lagrangian one-forms.
- The nonabelian version suggests a variational principle for any Hamiltonian group action; one could test whether known time-dependent invariants, such as Ermakov\u2013Lewis-type quantities, arise as moment-map components in some group parametrisation, which would extend the paper's time-dependent example.
- Because the phase-space multiforms of AKNS-type field theories share the same coadjoint-orbit structure, the same one-step reformulation may lift to $1+1$ field hierarchies; the authors only note this as an open question.
- A path-integral quantization over curves in $M\times \mathbb{R}^n$ using $\int \gamma^*L$ would automatically sum over multi-time surfaces, with on-shell closedness of $L$ making the result surface-independent; the paper mentions this direction but does not develop it.
Signed reviews
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. This paper introduces a phase-space reformulation of Lagrangian multiform theory for finite-dimensional systems. The central object is the one-form L = p_µ dq^µ - H_i dt^i on M × R^n, together with a univariational principle that requires every curve in an n-dimensional submanifold to be stationary. Section 2.2 derives the resulting equations (2.26)-(2.28) and shows that they combine the multi-time Euler-Lagrange equations with the closure relation, and that local solutions exist if and only if the Hamiltonians Poisson commute. Section 3 attempts to prove equivalence with conventional position-space Lagrangian one-forms through a generalized Legendre transform, claiming in the abstract and introduction that any finite-dimensional Lagrangian one-form, or any Liouville integrable system, can be recast in the new framework. Section 4 replaces R^n by a Lie group and shows that the Euler-Lagrange equations imply {H_i,H_j} = c^k_ij H_k, so that H is a moment map for a Hamiltonian group action. The proof of the universal recasting claim rests on the solvability of equation (3.4) and on the convexity and invertibility of a chosen combination α^i H_i in equation (3.7); these hypotheses are not shown to hold generally, and the paper's own Toda example and the velocity-linear discussion in Section 3.4 illustrate that the claimed universality goes beyond what is proved.
Significance. Section 2.2 is clean and checkable: substitution of (2.26)-(2.27) into (2.28) directly yields the Poisson commutator, and the Frobenius argument gives local existence of integral manifolds. The nonabelian extension in Section 4 is genuinely interesting and appears to give a new variational derivation of the moment-map condition for Hamiltonian group actions. However, the headline claim that every finite-dimensional Lagrangian one-form, or every Liouville integrable system, can be recast in the position-space framework is not supported by the proof. The inverse Legendre transform of Section 3.1 requires assumptions that fail in the paper's own Toda example for β ≠ 0, where the solution (3.33) is two-valued, and in the velocity-linear case treated only partially in Section 3.4. These gaps do not undermine the phase-space derivation or the Hamiltonian group-action result, but they are load-bearing for the advertised universality of the method.
major comments (3)
- [Abstract and Section 3.1, Eqs. (3.4), (3.7)] The claim that any finite-dimensional Lagrangian one-form, or any Liouville integrable system, can be recast in the phase-space framework is not established. The inverse Legendre transform requires solving the overdetermined system (3.4) for the velocities v^µ_j, and then requires that a vector α exist such that α^i H_i is a convex function of p_µ with invertible Hessian g_µν, so that equation (3.7) determines p_µ uniquely. No theorem is given showing that these conditions hold for arbitrary Lagrangian one-forms or arbitrary Liouville integrable systems; the text introduces them as assumptions immediately after (3.4). Section 3.4 then explicitly excludes the velocity-linear degenerate case. The abstract and introduction should either state these hypotheses precisely or the universality claim should be withdrawn.
- [Section 3.3, Eq. (3.33)] The Toda example contradicts the uniqueness premise of the inverse Legendre transform. For β ≠ 0 the combination α^1 H_1 + α^2 H_2 is cubic in p_µ and therefore not globally convex, and the displayed solution (3.33) is two-valued, parameterized by m sign choices. The paper presents this as a family of Lagrangian multiforms, but the unique-solution statement in Step (3) and the right-inverse calculation (3.11)-(3.16) do not apply in this case. The revision should either prove an extended theorem covering non-convex or branching cases, or explicitly restrict the equivalence theorem to the convex regime and present the Toda construction as a formal extension rather than as an instance of the general proof.
- [Section 3.4] The treatment of velocity-linear Lagrangian one-forms is incomplete in a way that directly affects the universality claim. The Lagrangian multiforms constructed in earlier work [5,6] for large classes of finite-dimensional integrable models are linear in velocities, so this is not a marginal case. The paper handles only the case where Ω in (3.40) is nondegenerate and can be brought to Darboux form by a point transformation; the degenerate multi-time case is explicitly declared 'much more complicated and beyond the scope of this article.' Thus the claimed systematic construction of position-space Lagrangian one-forms for any Liouville integrable system is not achieved for this class.
minor comments (3)
- [Eq. (2.37)] In the variation of the global action, the intermediate integral is written over Σ while ∆ is the surface defined in (2.36); this should be ∆ to avoid confusion.
- [Section 4.2] The statement that (4.13) holds for all Y^j because of 'similar arguments to those presented in section 2.2' would benefit from the explicit invertibility argument used in (2.33)-(2.35), since in the nonabelian case the submanifold is not a priori a graph over G.
- [Section 4.3, Eq. (4.35)] There are typographical errors in the displayed system: a double comma appears after the first brace, and the spacing is inconsistent. The affiliation line also contains 'Unite d Kingdom' and Section 4.3 contains 'framewoork'; these should be corrected.
Circularity Check
No significant circularity: the phase-space and Lie-group derivations are self-contained, and the inverse Legendre-transform caveats are scope limitations rather than circular reductions.
full rationale
The paper's central derivation is self-contained. Starting from the phase-space one-form L = p_mu dq^mu - H_i dt^i, the univariational equations (2.26)-(2.28) are obtained directly from demanding criticality for all curves in the hypersurface Sigma, and substituting (2.26)-(2.27) into (2.28) yields the Poisson commutator {H_j,H_i}=0 in (2.29). This is an output of the variational equations, not an input. The converse existence statement is proved by an independent Frobenius argument: the distribution defined by X lrcorner dL = 0 is integrable because dL is closed, and its rank is shown to be n using the Pfaffian system theta^mu, phi^mu and equation (2.32); Poisson commutativity is used explicitly to identify the kernel of the Pfaffian system with the distribution. In the nonabelian section, the Lagrangian one-form L = alpha - (H, g^{-1}dg) is chosen with an arbitrary smooth map H, and the Euler-Lagrange equations (4.10)-(4.12) are computed from dL using the Maurer-Cartan equation. Substituting the first two equations into the third gives (4.13), and the univariational requirement that this hold for all curves forces {H_i,H_j} = c^k_ij H_k in (4.14). Thus the moment-map condition is derived, not assumed; the paper explicitly stresses this point in Section 4.2 when distinguishing its starting point from the symplectic-quotient construction. The inverse Legendre transform in Section 3.1 does rest on unproved solvability of (3.4) and convexity/invertibility assumptions in (3.7), and Section 3.4 explicitly leaves the degenerate linear-velocity case open. These are genuine scope limitations and correctness risks, but they are not circular reductions: the paper does not fit parameters to the quantities it then claims to predict, and no equation is shown to be equivalent to its own input by construction. Self-citations to [5,6] and [22] are used as background, motivation, and comparison with earlier examples; the core derivations do not depend on those citations for their validity. The claimed variational origin of Hamiltonian group actions is a construction whose output condition coincides with the standard moment-map condition, but this is a normal Lagrangian-to-equation relationship rather than circularity.
Assumptions & free parameters
free parameters (1)
- alpha (with beta = alpha2/alpha1 in examples) =
arbitrary nonzero vector in R^n; no fitted value
assumptions (4)
- standard math Frobenius integrability theorem and Darboux theorem apply in the stated smooth category.
- domain assumption Boundary terms vanish for all variations, and the submanifold Sigma can be locally graphed over R^n.
- ad hoc to paper There exists alpha such that alpha^i H_i is convex in p_mu, and equation (3.4) can be solved for the velocities.
- domain assumption For velocity-linear Lagrangian one-forms, the induced two-form Omega = dp_mu wedge dq^mu is nondegenerate.
Cite this review
Pith. "Pith review of On the geometry of Lagrangian one-forms." pith.science (2026). https://pith.science/paper/7C2WMKQJ
@misc{pith2026241214700,
author = {Pith},
title = {Pith review of: On the geometry of Lagrangian one-forms},
year = {2026},
howpublished = {\url{https://pith.science/paper/7C2WMKQJ}},
note = {Machine review of arXiv:2412.14700}
}
read the original abstract
Lagrangian multiform theory is a variational framework for integrable systems. In this article we introduce a new formulation which is based on symplectic geometry and which treats position, momentum and time coordinates of a finite-dimensional integrable hierarchy on an equal footing. This formulation allows a streamlined one-step derivation of both the multi-time Euler-Lagrange equations and the closure relation (encoding integrability). We argue that any Lagrangian one-form for a finite-dimensional system can be recast in our new framework. This framework easily extends to non-commuting flows and we show that the equations characterising (infinitesimal) Hamiltonian Lie group actions are variational in character. We reinterpret these equations as a system of compatible non autonomous Hamiltonian equations.
Reference graph
Works this paper leans on
-
[22]
V. Caudrelier, F. Nijhoff, D. Sleigh, M. Vermeeren, Lagrangian multiforms on Lie groups and non-commuting flows , J. Geom. Phys. 187 (2023), 104807
work page 2023
- [1]
-
[2]
S. Yoo-Kong, S. Lobb, F. Nijhoff, Discrete-time Calogero-Moser system and Lagrangian 1-form structure, J. Phys. A44 (2011), 365203
work page 2011
-
[3]
Suris, Variational formulation of commuting Hamiltonian flows: mu lti-time Lagrangian 1-forms, J
Y.B. Suris, Variational formulation of commuting Hamiltonian flows: mu lti-time Lagrangian 1-forms, J. Geom. Mech. 5 (2013), 365
work page 2013
-
[4]
M. Petrera, Y.B. Suris, Variational symmetries and pluri-Lagrangian systems in cl assical mechanics, J. Nonlinear Math. Phys. 24 (2017), 121
work page 2017
-
[5]
V. Caudrelier, M. Dell‘Atti, A.A. Singh, Lagrangian multiforms on coadjoint orbits for finite-dimensional integrable systems , Lett. Math. Phys. 114 (2024), 34
work page 2024
-
[6]
V. Caudrelier, A.A. Singh, B. Vicedo, Lagrangian Multiform for Cyclotomic Gaudin Models , SIGMA 20 (2024), 100
work page 2024
-
[7]
P. Xenitidis, F. Nijhoff Frank, S. Lobb, On the Lagrangian formulation of multidimension- ally consistent systems , Proc. R. Soc. A. 467 (2011), 3295–3317
work page 2011
Show all 28 references
-
[8]
Richardson, M
J.J. Richardson, M. Vermeeren Discrete Lagrangian multiforms for quad equations, tetra- hedron equations, and octahedron equations , arXiv:2403.16845
-
[9]
Nijhoff, D.J
F.W. Nijhoff, D.J. Zhang, On the Lagrangian multiform structure of the extended latti ce Boussinesq system , Open Communications in Nonlinear Mathematical Physics, S pecial Is- sue in Memory of Decio Levi (February 15, 2024) ocnmp:12759
2024
-
[10]
Suris, M
Y.B. Suris, M. Vermeeren, On the Lagrangian structure of integrable hierarchies , in Ad- vances in discrete differential geometry, Editor A.I. Bobenk o, Springer Berlin Heidelberg, 2016, 347–378
2016
-
[11]
Sleigh, F
D. Sleigh, F. Nijhoff, V. Caudrelier, A variational approach to Lax representations, J. Geom. Phys. 142 (2019), 66–79
2019
-
[12]
Sleigh, F
D. Sleigh, F. Nijhoff, V. Caudrelier, Variational symmetries and Lagrangian multiforms , Lett. Math. Phys. 110 (2020), 805
2020
-
[13]
Caudrelier, M
V. Caudrelier, M. Stoppato, Hamiltonian multiform description of an integrable hierar chy, J. Math. Phys. 61 (2020), 123506, 25
2020
-
[14]
Caudrelier, M
V. Caudrelier, M. Stoppato, Multiform description of the AKNS hierarchy and classical r-matrix, J. Phys. A54 (2021), 235204, 43
2021
-
[15]
Petrera, M
M. Petrera, M. Vermeeren, Variational symmetries and pluri-Lagrangian structures f or integrable hierarchies of PDEs , Eur. J. Math. 7 (2021), 741–765
2021
-
[16]
Caudrelier, M
V. Caudrelier, M. Stoppato, B. Vicedo, Classical Yang-Baxter equation, Lagrangian multi- forms and ultralocal integrable hierarchies , Comm. Math. Phys. 405 (2024), 12
2024
-
[17]
Lobb, F.W
S. Lobb, F.W. Nijhoff, A Variational Principle for Discrete Integrable Systems , SIGMA 14 (2018), 041. 20
2018
-
[18]
F.W. Nijhoff, Lagrangian multiform structure of discrete and semi-discr ete KP systems , Open Communications in Nonlinear Mathematical Physics, Pr oceedings: OCNMP Confer- ence, Bad Ems (Germany), 23-29 June 2024 (July 9, 2024) ocnmp :13809
2024
-
[19]
Sleigh, F.W
D. Sleigh, F.W. Nijhoff, V. Caudrelier, Lagrangian multiforms for Kadomtsev-Petviashvili (KP) and the Gelfand-Dickey hierarchy , Int. Math. Res. Not. IMRN (2023), 1420–1460
2023
-
[20]
Nijhoff, Lagrangian 3-form structure for the Darboux system and the K P hierarchy, Lett
F.W. Nijhoff, Lagrangian 3-form structure for the Darboux system and the K P hierarchy, Lett. Math. Phys. 113 (2023), 27
2023
-
[21]
Sleigh, M
D. Sleigh, M. Vermeeren, Semi-discrete Lagrangian 2-forms and the Toda hierarchy , J. Phys. A55 (2022), 475204
2022
-
[23]
Puttarprom, W
C. Puttarprom, W. Piensuk, S. Yoo-Kong, Integrable Hamiltonian Hierarchies and La- grangian 1-Forms, arXiv:1904.00582
1904 arXiv
-
[24]
Faddeev and R
L. Faddeev and R. Jackiw, Hamiltonian Reduction of Unconstrained and Constrained Sys - tems, Phys. Rev. Lett. 60 (1988), 1692
1988
-
[25]
Magazev, V.V
A.A. Magazev, V.V. Mikheyev, I.V. Shirokov, Computation of Composition Functions and Invariant Vector Fields in Terms of Structure Constants of Ass ociated Lie Algebras, SIGMA 11 (2015), 066
2015
-
[26]
Semenov-Tian-Shanskii, What is a classical r-matrix?, Funct
M.A. Semenov-Tian-Shanskii, What is a classical r-matrix?, Funct. Anal. Its Appl. 17 (1983), 259–272
1983
-
[27]
King, F.W
S.D. King, F.W. Nijhoff, Quantum variational principle and quantum multiform struc ture: The case of quadratic Lagrangians , Nucl. Phys. B 947 (2019), 114686
2019
-
[28]
Kongkoom, S
T. Kongkoom, S. Yoo-Kong, Quantum integrability: Lagrangian 1-form case , Nucl. Phys. B987 (2023), 116101. 21
2023
Reviewed August 11, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.