A spectral theorem for matrices over fields of power series (Q1898751)

The authors consider matrices over fields \(K = \mathbb{R} ((t_1, t_2, \dots, t_m))\) of formal power series in one or several variables with real coefficients. The main result states that over these \(K\) every symmetric matrix can be orthogonally diagonalized. It is also shown that this result even carries over to fields of generalized power series in infinitely many variables. The method of proof can be outlined as follows. First write \(K = K_0 ((t))\) where \(t = t_m\) and \(K_0 = \mathbb{R} ((t_1, \dots, t_{m - 1}))\). The field \(K = K_0 ((t))\) is complete with respect to a nonarchimedean discrete valuation. This allows to present a given symmetric matrix \({\mathcal A} \in \text{Mat}_n (K)\) as a power series \({\mathcal A} = A_0 + A_1t + A_2 t^2 + \dots\) with coefficients \(A_k \in \text{Mat}_n (K_0)\). Secondly a recursive construction is set up that produces an orthogonal transition matrix \({\mathcal U} = U_0 + U_1t + U_2t^2 + \dots\) such that \({\mathcal U}^{tr} {\mathcal A} {\mathcal U}\) is decomposed into two blocks of smaller size. The proof is then finished by an induction. The eigenvalues of \({\mathcal A}\) are obtained at the end as a by-product.