Simple transcendental extensions of valued fields. II: A fundamental inequality (Q1068896)

Let \(K_ 0\subset K=K_ 0(x)\) be fields with x transcendental over \(K_ 0:\) let \(v_ 0\) be a valuation of \(K_ 0\) and v be an extension of \(v_ 0\) to K; and let \(V_ 0\subset V\), \(k_ 0\subset k\), and \(G_ 0\subset G\) be the respective valuation rings, residue fields, and value groups. This paper is concerned with extensions such that \(k/k_ 0\) is not algebraic. For such extensions there exists \(t\in V\) such that t specializes to a \(t^*\) transcendental over \(k_ 0.\) For any \(t\in K\) and \(\not\in K_ 0\), define deg t to be [K: \(K_ 0(t)]\). Let \(E= least\) n such that there exists \(t\in V\) of deg n such that t specializes to a transcendental over \(k_ 0\); let \(I=[G: G_ 0]\); and let \(R=[k_ 0': k_ 0]\), where \(k_ 0'\) denotes the algebraic closure of \(k_ 0\) in k. It is easily seen that \(E\geq I R\), and this is the fundamental inequality of the title. Conjecture: Suppose rk \(v_ 0=1\). If either \(v_ 0\) is discrete or char \(k_ 0=0\), then \(E=I R\). This conjecture is verified here for the case that \(I=1.\) Recently (May, 1986) \textit{M. Matignon} has formulated and proved a natural generalization of this conjecture; it will appear in a joint publication with the reviewer. In addition, Matignon has pointed out that the discrete case of the conjecture was proved in the unpublished 1968 Saarbrücken dissertation of \textit{H. Mathieu}. [For part I, see ibid. 22, 201-221 (1982; Zbl 0501.12022).]