Moser's Stability Theorem
Let be a compact manifold and , , a smooth family of symplectic forms on with for all .
Then there exists a family of diffeomorphisms with such that:
That is, all symplectic forms in the same cohomology class are equivalent via diffeomorphism.
This profound result, proved by JΓΌrgen Moser in 1965, shows that symplectic structures are stable under deformation within a fixed cohomology class. It extends Darboux's theorem from local to global settings (on compact manifolds).
Geometric Interpretation: If we smoothly deform a symplectic form without changing its cohomology class, we can "track" this deformation by a family of diffeomorphisms. The symplectic structure itself doesn't change β only our coordinate description.
The compactness assumption is essential. On non-compact manifolds, diffeomorphisms may "escape to infinity" and the theorem fails. For example, on , scaling changes cohomology in a relative sense.
The proof uses Moser's path method: solve the equation for a time-dependent vector field , then integrate to obtain .
The condition constant in is necessary: if two symplectic forms define different cohomology classes, they cannot be related by a diffeomorphism, since:
for any diffeomorphism .
On a compact KΓ€hler manifold , the KΓ€hler form lies in a specific KΓ€hler cohomology class. Moser's theorem implies that small deformations of within this class are related by diffeomorphisms, even though the complex structure may change.
This is crucial in complex geometry: the space of KΓ€hler structures modulo diffeomorphism is finite-dimensional (determined by Hodge theory).
Moser's theorem implies that for a compact symplectic manifold , the space of symplectic structures in is:
The quotient is a single point! This shows remarkable rigidity: fixing cohomology class determines the symplectic structure up to diffeomorphism.
Consider the 2-torus with symplectic forms:
for . These have different cohomology classes: .
By Moser, and are diffeomorphic if and only if . The area (total symplectic volume) is the only invariant.
Moser's theorem contrasts with Riemannian geometry: Riemannian metrics in the same conformal class are NOT generally related by diffeomorphisms. The Yamabe problem asks when metrics of constant scalar curvature exist in a conformal class β a non-trivial question.
Symplectic geometry is simpler: cohomology class alone determines the structure.
Moser's theorem is crucial for symplectic reduction (Marsden-Weinstein): when a Lie group acts on with moment map , the reduced space:
inherits a symplectic structure. Moser's method proves that this structure is well-defined independent of choices.
In Hamiltonian mechanics, perturbations of the symplectic structure with can be "absorbed" by coordinate changes. This means:
Physical perturbations (changing or in a cohomologically trivial way) are equivalent to coordinate changes in the unperturbed system.
Only cohomologically non-trivial perturbations genuinely change the dynamics.
On non-compact manifolds, variants of Moser's theorem hold:
- Proper support: If differ only on a compact set, Moser applies
- Bounded geometry: Control on derivatives allows Moser on complete manifolds
- Relative version: Fix behavior at infinity, apply Moser to interior
These extensions are crucial for local-to-global principles in symplectic topology.
Moser's stability has analogues in other geometries:
- Contact geometry: Gray's stability theorem for contact forms
- Complex geometry: Kodaira-Spencer deformation theory
- Riemannian geometry: No direct analogue (metrics not determined by cohomology)
The symplectic case is distinguished by its simplicity.