...
ProvedTeichmuller.exists_teichDist_T_ltaether-cataloggeometry
... but the infimum of the parabolic displacement is zero: it is not attained. This is the cusp of the moduli space of tori, and the reason the moduli space is not compact.
theorem Teichmuller.exists_teichDist_T_lt{ε : ℝ} (hε : 0 < ε) :
∃ σ : ℍ, teichDist σ (ModularGroup.T • σ) < ε := by sorry
Formalization Note Transplanted verbatim from the Aether Catalog source Geometry/Teichmuller/ModuliSpace.lean; the statement is byte-identical to the source declaration, elaborated with autoImplicit disabled in the platform environment.
Preamble
-- Thm stub generated from Geometry/Teichmuller/ModuliSpace.lean
import Mathlib
import Definitions.Def_Geometry_Teichmuller_ModuliSpace
import Definitions.Def_Geometry_Teichmuller_TorusSpace
import Definitions.Def_Geometry_Teichmuller_TranslationLength
/-
# The mapping class group, the moduli space of tori, and its degeneracies
The mapping class group of the torus is `SL(2, ℤ)`, acting on the Teichmüller space `ℍ` of
marked tori by change of marking, i.e. by Möbius transformations. The moduli space of tori is
the quotient `ℍ / SL(2, ℤ)`, whose (pseudo-)distance is
d_M (τ, τ') = inf over the mapping class group of d_T (τ, g · τ') .
This file proves:
* `Teichmuller.teichDist_smul` : **the mapping class group acts by isometries** of the
Teichmüller metric — the property that makes the moduli distance well defined;
* `Teichmuller.moduliDist_comm`, `moduliDist_triangle`, `moduliDist_nonneg`,
`moduliDist_smul_left`, `moduliDist_smul_right` : the moduli distance is a
mapping-class-group-invariant pseudometric on `ℍ`, i.e. a genuine distance on the quotient;
* `Teichmuller.moduliDist_self_smul` and `Teichmuller.exists_ne_moduliDist_eq_zero` :
the moduli distance is *only* a pseudometric on `ℍ`: distinct marked tori can be
isomorphic as unmarked tori — the quotient is nontrivial;
* `Teichmuller.exists_nontrivial_stabilizer` : **the action is not free**: the square root of
`-1` in `SL(2, ℤ)` fixes the square torus `i` and is nontrivial even modulo `±1`; hence the
moduli space is an orbifold, not a manifold, and the quotient map is not a covering;
* `Teichmuller.exists_order_three_stabilizer` and `Teichmuller.smul_rho_ne_I` : the hexagonal
torus `ρ = -1/2 + i√3/2` carries a stabiliser of order **three** in `PSL(2, ℤ)`, and
`Teichmuller.sq_eq_one_of_smul_I_eq` shows the stabiliser of `i` has order **two**; hence the
two cone points lie in different orbits and the moduli space has at least two distinct
orbifold singularities, of cone angles `π` and `2π/3`;
* `Teichmuller.teichDist_T_pos` together with `Teichmuller.exists_teichDist_T_lt` :
the parabolic mapping class `T : τ ↦ τ + 1` has **zero translation length which is not
attained** — the quantitative source of the noncompactness of the moduli space
(its "cusp"), in sharp contrast with the Anosov classes of
`Geometry.Teichmuller.TranslationLength`, whose translation length is positive and attained.
-- !-- Lab Notes -- !--
Hypothesizer: the three Nielsen–Thurston types (elliptic / parabolic / Anosov) should be
*visible in the metric* on Teichmüller space: fixed point, infimum-zero-not-attained, and
positive-attained respectively.
Experimenter: the displacement identity `cosh_dist_smul` computes all three at once. For
`S = !![0,-1;1,0]` (trace `0`) the displacement vanishes at `i`; for `T = !![1,1;0,1]`
(trace `2`) it equals `arcosh (1 + 1/(2y²)) > 0`, tending to `0` as `y → ∞`; for trace `> 2`
it is bounded below by `arcosh ((t²-2)/2) > 0`. Analyst: the three cases are exactly
`|tr| < 2`, `= 2`, `> 2` — the metric geometry of the moduli space is governed by the trace,
and the failure of properness at the cusp is the parabolic case.
Critic: "the action is not free" is weaker than "the orbifold has two distinct cone points" —
the latter needs the stabiliser of `i` computed exactly. Experimenter: solving `g · i = i`
over `ℤ` forces `a = d`, `b = -c`, hence `a² + c² = det g = 1`, whose only integral solutions
are the four points `(±1, 0), (0, ±1)`; all four square to `±1`. Conjugating the order-three
stabiliser of `ρ` into the stabiliser of `i` therefore yields a contradiction, so the two cone
points are distinct in moduli.
-/
open Teichmuller
open Complex UpperHalfPlane Matrix MatrixGroups
variable (τ τ' τ'' : ℍ)Formal statement
theorem Teichmuller.exists_teichDist_T_lt{ε : ℝ} (hε : 0 < ε) :
∃ σ : ℍ, teichDist σ (ModularGroup.T • σ) < ε := by sorrySource