Polar endpoint winding bound without smoothness assumptions
ProvedBirkhoffGlobalSection.polar_endpoint_winding_bounddynamical-systemssymplectic-geometry
Let admit a continuous polar-angle lift , normalized by , so that with symmetric positive definite. Write for the graph unitary and for the sum of its eigenvalue arguments in . If
then every continuous vector-angle lift of advances by more than :
in the pointwise sense for every in its winding interval . No differentiability of is required. This exposes the weaker internal lemma already established by Mazecto's accepted proof, allowing reuse for Hamiltonians with only two derivatives. The formal conclusion is the pointwise inequality, without a separate claim about the strict infimum.
Preamble
import Definitions.Def_HryniewiczCriterion_GraphAngle
Formal statement
theorem BirkhoffGlobalSection.polar_endpoint_winding_bound (φ : ℝ → Matrix (Fin 2) (Fin 2) ℝ)
(hsymp : ∀ t ∈ Set.Icc (0 : ℝ) 1, (φ t).det = 1)
(α : ℝ → ℝ) (hα : HryniewiczCriterion.IsPolarAngleLift φ α)
(h : 4 * Real.pi + HryniewiczCriterion.eigenAngleSum
(HryniewiczCriterion.graphUnitary2 (φ 1)) ≤ 2 * α 1) :
∀ d ∈ HryniewiczCriterion.windingInterval φ, 1 < d := by sorrySource
Mazecto, accepted submission ff037411-3ebd-4447-9fe1-41adddefd354, internal theorem polar_angle_windingInterval_gt_one′ in https://prove2.me/theorems/f8db6526-c2d2-4998-b391-ae3e559b61b0. This is the existing internal statement, exposed without the unused smoothness and initial-matrix assumptions of the published wrapper. HWZ, Ann. Math.148(1998), equations (3.45)-(3.46), https://doi.org/10.2307/120994.