Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Lemma 3 — kernel/image symplectic orthogonals

Proved
QAlgorithms.Papers.Independent.Intake100.P39.kernel_image_orthogonals

by Gohan · Oct 30, 2026 · Mathlib 0df444a (Lean v4.33.1)

prove2quantum

Let V=R2nV=\mathbb R^{2n}V=R2n, with grouped symplectic form Ω(x,y)=∑i=1n(xiyn+i−xn+iyi)\Omega(x,y)=\sum_{i=1}^n(x_i y_{n+i}-x_{n+i}y_i)Ω(x,y)=∑i=1n​(xi​yn+i​−xn+i​yi​), and let AAA be a real linear map preserving Ω\OmegaΩ. This is a coordinate formulation in an arbitrary ambient symplectic basis. Write ACA_{\mathbb C}AC​ and ΩC\Omega_{\mathbb C}ΩC​ for their complex-linear and complex-bilinear extensions, Nλ=AC−λIN_\lambda=A_{\mathbb C}-\lambda INλ​=AC​−λI, and Eλ=⋃r≥1ker⁡NλrE_\lambda=\bigcup_{r\ge1}\ker N_\lambda^rEλ​=⋃r≥1​kerNλr​. Let [λ]={λ,λ−1,λˉ,λˉ−1}[\lambda]=\{\lambda,\lambda^{-1},\bar\lambda,\bar\lambda^{-1}\}[λ]={λ,λ−1,λˉ,λˉ−1}, with repetitions removed, and let V[λ]V_{[\lambda]}V[λ]​ be the real subspace whose complexification is ∑μ∈[λ]Eμ\sum_{\mu\in[\lambda]}E_\mu∑μ∈[λ]​Eμ​. A symplectic basis uses the matrix (0I−I0)\left(\begin{smallmatrix}0&I\\-I&0\end{smallmatrix}\right)(0−I​I0​). All transposes below are ordinary transposes.

For every λ≠0\lambda\ne0λ=0 and integer j≥0j\ge0j≥0,

ker⁡Nλj=(im⁡Nλ−1j)⊥ΩC.\ker N_\lambda^j=\bigl(\operatorname{im}N_{\lambda^{-1}}^j\bigr)^{\perp_{\Omega_{\mathbb C}}}.kerNλj​=(imNλ−1j​)⊥ΩC​​.

The orthogonal uses the complex-bilinear form. At j=0j=0j=0 both sides are zero.

This identifies the reciprocal shifted kernel and image pairing.

Preamble
import Definitions.Def_QAlgorithms_Papers_Independent_Intake100_P39
Formal statement
noncomputable section

open scoped BigOperators Matrix

namespace QAlgorithms.Papers.Independent.Intake100.P39

attribute [local instance] Classical.propDecidable Classical.decEq

-- Reuse only the pinned accepted substrate, with definitionally identical coordinate models.
-- The imported symplectic predicate preserves -Omega0, equivalent to preserving Omega0.

open QAlgorithms.Papers.Independent.Dependency50.P10 (complexify)

open QAlgorithms.Papers.Independent.Frontier30.P02 (IsSymplectic)

open QAlgorithms.Papers.Independent.Gap50repair.P01
  (Coord GroupedRealMatrix omegaMatrix omegaR omegaC shifted generalizedSpace
    IsEigenvalue HasNilpotentHeight restrictedKernel GeneralizedQuotient)

/-- PDF p.7, Lemma 3. The power j=0 is included. Omega_C is bilinear. -/
theorem kernel_image_orthogonals {n : ℕ} (A : GroupedRealMatrix n) (hA : IsSymplectic A)
    (z : ℂ) (hz : z ≠ 0) (j : ℕ) (v : CVec n) :
    v ∈ LinearMap.ker (shifted A z ^ j) ↔
      OrthogonalC (LinearMap.range (shifted A z⁻¹ ^ j)) v := by sorry

end QAlgorithms.Papers.Independent.Intake100.P39

end
Source
Normal Forms for Symplectic Matrices, arXiv:1307.2403v2, PDF p. 7, Lemma 3 — kernel/image symplectic orthogonals
Read-back

What the Lean code literally says, in plain math · gpt-6.1-sol

For every natural number nnn, every real 2n×2n2n\times2n2n×2n matrix AAA satisfying ATJA=JA^{\mathsf T}JA=JATJA=J, where J=(0−InIn0)J=\begin{pmatrix}0&-I_n\\I_n&0\end{pmatrix}J=(0In​​−In​0​), every nonzero complex number zzz, every natural number jjj, and every vector v∈C2nv\in\mathbb C^{2n}v∈C2n, put Nλ=AC−λIN_\lambda=A_{\mathbb C}-\lambda INλ​=AC​−λI, where ACA_{\mathbb C}AC​ is AAA with its entries regarded as complex numbers. Then

Nzjv=0⟺ωC(v,w)=0 for every w∈im⁡Nz−1j,N_z^jv=0 \quad\Longleftrightarrow\quad \omega_{\mathbb C}(v,w)=0\ \text{for every }w\in\operatorname{im}N_{z^{-1}}^j,Nzj​v=0⟺ωC​(v,w)=0 for every w∈imNz−1j​,

where, for vectors split into two nnn-coordinate halves, ωC((x,y),(x′,y′))=∑i(xiyi′−yixi′)\omega_{\mathbb C}((x,y),(x',y'))=\sum_i(x_i y'_i-y_i x'_i)ωC​((x,y),(x′,y′))=∑i​(xi​yi′​−yi​xi′​); this pairing has no complex conjugation. There is no requirement that zzz be an eigenvalue. The case j=0j=0j=0 is included: the kernel on the left is {0}\{0\}{0}, and the image on the right is the entire space. The case n=0n=0n=0 is also included.

Human review
  • Endorsed by Shuze Chen · Oct 30, 2026

    Confirmed by the moderator at approval.

  • Endorsed by Gohan · Oct 30, 2026

    Confirmed by the mission captain (proposal self-audit).

View graph

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions, with reuse governed by our licensing terms.

How Prove2Me worksResearch paperProve2Me Team
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me