Location of zeros. II: Ordered fields (Q1171090)

This paper is the second part of an investigation devoted to multiplier sequences and n-sequences. The definitions and notations used in Part I for the field of real numbers, are carried over to arbitrary fields. In section 2 the authors show that Polya and Schur's characterization of multiplier sequences can be proved only for a special class of ordered fields. In particular the following result holds: Let F be an ordered field and \(F^*\) denote the intersection of all real closures of F inside a fixed algebraic closure of F. Let \(\Gamma\) denote a sequence of zeros and totally positive elements of F then the following statements are equivalent: a) if \(\Gamma [(x+1)^ n]\) splits in \(F^*\) then \(\Gamma\) is an n-sequence, b) \(F=F^*.\) In the third section multiplier sequences and n-sequences \(\Gamma\) which satisfy the fundamental inequality \(Z_ c[\Gamma(f)]\leq Z_ c(f)\) are studied (for a polynomial f with coefficients in F, \(Z_ c(f)\) denotes the number of roots of f which do not lie in F). It turns out that for every real closed subfield of the field of real numbers, every multiplier sequence satisfies the fundamental inequality, but the authors do not know whether this theorem is true in general. In the paper some multiplier sequences and n-sequences which satisfy the fundamental inequality for all real closed fields are constructed. Section 4 contains some partial result concerned with fields which satisfy Rolle's theorem and its generalization. A complete characterization of fields for which Rolle's theorem holds has been given by \textit{R. Brown}, \textit{T. Craven}, and \textit{M. J. Pelling} [Ordered fields satisfying Rolle's theorem, preprint]. Multiplier sequences in arbitrary fields are studied in the fifth section. In particular it is obtained an interesting characterization of algebraically closed fields. Moreover the existence of inverses for elements in the semigroup of multiplier sequences is examined. It is proved that if F is an algebraic extension of a finite field then there exists a so-called generalized inverse element for every multiplier sequence.