Machine-checked, not peer-reviewed; no priority claim.

Sequential products & rigidity / 03

Classification of sequential products

Real matrix effects have the Lüders form; complex matrix effects of size at least three have a unique twist parameter.

The statement

On the real Hermitian matrices of size N ≥ 1, a sequential product satisfying S1–S7 agrees with the Lüders formula on effects. On the complex Hermitian matrices of size N ≥ 3, it has the twist form for a unique real parameter t. The complex conclusion holds for every pair of effects, including singular ones.

Real: a & b = √a · b · √a.

Complex: a & b = a¹ᐟ²⁺ⁱᵗ · b · a¹ᐟ²⁻ⁱᵗ.

An effect is a Hermitian matrix in the Loewner interval 0 ≤ a ≤ 1. The complex powers are defined by continuous functional calculus, with the twist factor taking value zero at the zero eigenvalue.

Hypotheses

  • A supplied SeqProd N ℝ or SeqProd N ℂ carries closure on effects and the algebraic axioms S1 and S3–S7 of van de Wetering.
  • The separate hypothesis S2 P requires continuity in the first argument for each fixed effect in the second argument, measured in the order-unit norm.
  • The real theorem assumes 0 < N; the complex theorem assumes 3 ≤ N. The equations concern effect pairs only.

The full axiom definitions are in the statement file. S2 is explicit in both theorem declarations; it isn’t hidden in SeqProd.

What’s not proved

Both conclusions are necessity statements for a supplied product. The unique-existence quantifier gives a unique parameter for that product; it doesn’t assert that products exist.

Neither converse is registered in this entry. The repository proves complex-family admissibility elsewhere; it doesn’t construct the real Lüders product as an admissible SeqProd satisfying S2.

This entry doesn’t classify complex 2 × 2 products, quaternionic or exceptional Jordan cases, or all simple Euclidean Jordan algebras. Values of the total operation off the effect interval aren’t constrained by the conclusion.

In Lean

Declaration excerpts from Solution.lean, in the namespace TwistNormalForm. Proof bodies are omitted; the source contains the definitions and proofs.

Ambient context
open scoped Matrix ComplexOrder
real_classificationSource
theorem real_classification {N : ℕ} (hN : 0 < N)
    (P : SeqProd N ℝ) (hS2 : S2 P)
    {a b : Mat N ℝ} (ha : IsEffect a) (hb : IsEffect b) :
    P.sp a b = luders a b
complex_classificationSource
theorem complex_classification {N : ℕ} (hN : 3 ≤ N)
    (P : SeqProd N ℂ) (hS2 : S2 P) :
    ∃! t : ℝ, ∀ a b : Mat N ℂ, IsEffect a → IsEffect b →
      P.sp a b = twist t a b
The registered source revision

Source & registry

Repository
ehrlich-b/sequential-products-lean
Commit
9c6d92f5fc22
Statement
Challenge.lean
Proof
Solution.lean
Configuration
comparator.json
Axiom audit
AxiomAudit.lean
Palomar
PALOMAR-2026-08-22-000006
Theorems
  • TwistNormalForm.real_classification
  • TwistNormalForm.complex_classification

The challenge files contain placeholders for the problems. The solution files discharge them. The registered commit supplies the excerpts and source links on this page.