LeanProofs.Crypto.GLEquivariance
DefinitionGrade4_LeanProofs_Crypto_GLEquivarianceclifford-algebraleanproofs-grade4rosetta
Foundational definitions and structural lemmas from LeanProofs.Crypto.GLEquivariance, required by the selected LeanProofs theorem graph.
Definition code
/-
Equivariance of the symbolic Navier-Stokes RHS under conjugation by
any invertible element — the GL generalization of Equivariance.lean.
## What changed and why
`Equivariance.lean` proved
rotorConj R (pdeRHS^[k] F) = pdeRHS^[k] (rotorConj R F)
under `hR : IsRotor60 R`. Tracing that proof shows:
- `rotorConj_add`, `rotorConj_smul` — no hypothesis at all
- `ConstantRotorDerivation.comm_one` — `IsRotor60` appears only in the
*type* of the axiom; the axiom has no internal proof that uses it
- `rotorConj_comm` → `clConj_preserves_comm` — inspecting the body:
the parameter `_hR : R_el * R_inv = 1` is marked unused (`_`);
only `hRi : R_inv * R_el = 1` is rewritten
Therefore the proof uses NO rotor-specific property (even grade, spinor
norm = 1). The only invertibility condition consumed is `M_inv * M = 1`.
Replacing `IsRotor60` with that single hypothesis and `rev60` with a
free `M_inv` yields the GL generalization with an identical proof skeleton.
## Practical consequence
Option 3 non-abelian signatures (conjugacy search in GL(64, 𝔽_p)) rest
on the same equivariance guarantee as the rotor protocol. The algebraic
soundness is machine-checked here; hardness of conjugacy search remains
a cryptanalytic assumption.
## Scope
MACHINE CHECKED here:
`conjBy M M_inv (pdeRHSGL^[k] F) = pdeRHSGL^[k] (conjBy M M_inv F)`
for all M M_inv : Cl60 satisfying only `M_inv * M = 1`.
ASSUMED (same as Equivariance.lean):
- Continuous PDE flow inherits equivariance in the dt → 0 limit.
- Hardness of vectorization / conjugacy search.
-/
import Definitions.Def_Grade4_LeanProofs_Crypto_Equivariance
namespace LeanProofsGrade4P2M
namespace Cl60Conj
end Cl60Conj
namespace ConstantInvertibleDerivation
end ConstantInvertibleDerivation
end LeanProofsGrade4P2M
Source