Abstract
The Weyl calculus for a pair A = (A1, A2) of self-adjoint (n × n)-matrices, due to H. Weyl, associates a matrix W A (ƒ) to each smooth function ƒ defined on R2 in a linear but typically not multiplicative way. Letting c A(λ) := det((A1 - λ1I)2 + (A2 - λ2I)2) for λ ∈ R2 denote the joint characteristic polynomial of the pair A, it is known, for n ≤ 3, that A1A2 = A2A1 if and only if WA(cA) = 0. It is an open question whether this is still true for n ≥ 4. Our aim here is to pursue two new approaches: the role of the canonical order structure for self-adjoint matrices; and topological invariants arising from continuity properties of the non-linear map (A, ƒ) → WA(ƒ).[Full Text (PDF)]