The Symplectic Slice Theorem

Juan-Pablo Ortega, Tudor S. RaţiuBirkhäuser 2004 · DOI · arXiv · OpenAlex

Abstract

Most of the good technical behavior of proper Lie group actions is a direct consequence of the existence of slices and tubes; they provide a privileged system of semiglobal coordinates in which the group action takes on a particularly simple form. Proper symplectic Lie group actions turn out to behave similarly: the tubular chart can be constructed in such a way that the expression of the symplectic form is very natural and, moreover, if there is a momentum map associated to this canonical action, this construction provides a normal form for it. The statement and proof of this Symplectic Slice Theorem is the main goal of this chapter.

Introduction

In the study of symplectic geometry, understanding how symmetry groups—specifically Lie groups—act on manifolds is fundamental. When these actions preserve the underlying symplectic structure, they are termed "canonical." A cornerstone of this field is the Symplectic Slice Theorem, originally formulated by Marle, Guillemin, and Sternberg (MGS). This theorem provides a "normal form," or a standardized local model, for the neighborhood of a group orbit, effectively simplifying complex global dynamics into manageable local descriptions.

However, the classical MGS theorem carries a significant constraint: it requires the existence of a global momentum map. While momentum maps elegantly link symmetries to conserved quantities (analogous to the relationship between rotational symmetry and angular momentum), they do not exist for every canonical action. Ortega and Ratiu address this limitation by generalizing the theorem to all proper canonical Lie group actions, regardless of whether a global momentum map is present.

Methodology: Beyond the Momentum Map

The authors' primary innovation is the replacement of the momentum map with the Chu map. Unlike its predecessor, a Chu map—which maps to Lie algebra two-cocycles—is guaranteed to exist for any canonical action. This allows for a universal geometric framework.

The construction follows several rigorous steps:

  1. The Symplectic Twist Product: The authors define a model space, $Y_r$, using a "twist product" construction. This space serves as the local template for the manifold's geometry near a specific orbit.
  2. Symplectic Reduction: Utilizing the Marsden-Weinstein reduction technique, they demonstrate that this model space $Y_r$ is a well-defined symplectic manifold.
  3. Equivariant Symplectomorphism: The core of the proof lies in establishing a transformation that preserves both the group action and the symplectic structure, mapping the original manifold's neighborhood to the model space $Y_r$.
  4. Reconstruction Equations: To bridge the gap between the model and the original dynamics, the authors derive differential equations (Hamilton’s equations) within these "tubular" coordinates.

Key Insights

The Chu Map as a Universal Tool

The realization that the Chu map contains sufficient geometric information to reconstruct a local symplectic structure is a pivotal insight. It effectively "stands in" for the momentum map, providing the necessary data to build the normal form in cases where classical methods fail.

Formal Invariance of Dynamics

A surprising finding in the paper is that the "reconstruction equations"—which describe how to lift motion from the simplified model back to the full manifold—remain formally identical to those used in globally Hamiltonian systems. This suggests that much of the existing dynamical intuition for Hamiltonian systems can be directly applied to this broader class of actions.

Tubewise Hamiltonian Actions

The paper introduces the concept of "tubewise Hamiltonian" actions. These are actions that may not be Hamiltonian globally but behave as such within local invariant neighborhoods (tubes). This provides a more nuanced classification for group actions in symplectic geometry.

Results and Generalization

The primary result, Theorem 2.2, proves that for any symplectic manifold with a proper canonical Lie group action, every orbit possesses a $G$-invariant neighborhood symplectomorphic to the model space $Y_r$.

Crucially, this work is a true generalization:

  • If a global momentum map exists, the $Y_r$ model reduces precisely to the classical Marle-Guillemin-Sternberg normal form.
  • The authors provide specific conditions (Corollary 4.4) under which an action is guaranteed to be tubewise Hamiltonian, such as when the first cohomology group of the Lie group vanishes ($H^1(G)=0$) or when the orbit is isotropic.

Conclusion and Future Directions

Ortega and Ratiu's work provides a robust, universal normal form for proper canonical actions. While the framework is comprehensive for standard symplectic manifolds, the authors identify potential extensions into non-Abelian quasi-Hamiltonian spaces. Currently, the definitions in that subfield are incompatible with standard symplecticity, marking a clear path for future research. Although the presentation assumes a high degree of familiarity with 1980s-era classical proofs, it serves as a vital bridge for experts seeking to unify the treatment of Hamiltonian and non-Hamiltonian canonical actions.

Summary updated

Abstract

Most of the good technical behavior of proper Lie group actions is a direct consequence of the existence of slices and tubes; they provide a privileged system of semiglobal coordinates in which the group action takes on a particularly simple form. Proper symplectic Lie group actions turn out to behave similarly: the tubular chart can be constructed in such a way that the expression of the symplectic form is very natural and, moreover, if there is a momentum map associated to this canonical action, this construction provides a normal form for it. The statement and proof of this Symplectic Slice Theorem is the main goal of this chapter.
BibTeX
@inproceedings{ortega2004symplectic,
  author = {Juan-Pablo Ortega and Tudor S. Raţiu},
  title = {{The Symplectic Slice Theorem}},
  booktitle = {Birkhäuser},
  year = {2004},
  doi = {10.1007/978-1-4757-3811-7\_7},
}
Figure