Finite splittings of differential matrix algebras (Q6052906)

Let \(K\) be a differential field of characteristic zero and let \(S\) be a finite subset of \(K - \{0\}\). The authors call a differential field extesnion \(E\) of \(K\) a \emph{splitting field} of \(S\) if for every \(a \in S\) there is a \(y \in E\) with \(y^\prime=ay\). Let \(\mathcal A\) denote the \(\mathbb Q\) span of the set of logarithmic derivatives in \(K\), let \(\langle S \rangle\) denote the \(\mathbb Z\) span of \(S\), let \(\mathcal A_S=\mathcal A \cap \langle S \rangle\) and let \(\mathcal A^\prime_S\) denote any \(\mathbb Z\) complement to \(\mathcal A_S\) in \(\langle S \rangle\). The authors prove that if \(\mathcal A^\prime_S=0\) then there is a splitting field \(E\) for \(S\) which is finite dimensional over \(K\). Conversely, if \(\mathcal A^\prime_S \neq 0\) then any splitting field \(E\) for \(S\) must have transcendence degree at least the \(\mathbb Z\) rank of \(\mathcal A^\prime_S\), and that there exists a splitting field of transcendence degree equal to this rank. The authors apply these results to differential matrix algebras over the differential field \(K\): Such a differential matrix algebra is called split if the derivation on the matrices is the derivation of \(K\) applied coordinatewise. Any differential matrix algebra's derivation is this coordinatewise derivation plus an inner derivation by a matrix. For an arbitrary differential matrix algebra over \(K\), a differential field extension \(E\) is called a splitting field if the differential matrix algebra tensored over \(E\) becomes split over \(E\). The authors apply their results on splitting fields for subsets to show that, for the case \(K=F(t)\) where the derivation is zero on \(F\) and sends \(t\) to \(t\), that the transcendence degree of a splitting field for an differential matrix algebra can be calculated from the properties over \(\mathbb Q\) of the eigenvalues of the matrix giving the derivation on the algebra. (Note: some of the authors' results depend on Proposition 2 of their reference 2. Yidi Wang has recently pointed out that that `only if' part of that Proposition only holds under further conditions on the differential field.)