The polynomial closure is not topological (Q2154265): Difference between revisions

Let \(D\) be a commutative integral domain with quotient field \(K\). The \(D\)-polynomial closure of a subset \(S\) of \(K\) is defined as the largest subset \(\overline S\) of \(K\) for which \(\mathrm{Int}(S, D) =\mathrm{Int}(\overline S, D)\), where \(\mathrm{Int}(S, D) = \{f \in K[X] \mid f(S) \subseteq D\}\) is the ring of integer-valued polynomials over \(S\). Chabert proved that the \(D\)-polynomial closure is topological (that is, it is equal to the closure operator for some topology on \(D\)) if \(D\) is a valuation domain of rank \(1\), using pseudo-convergent sequences introduced by Ostrowski. The authors prove that the \(V\)-polynomial closure is never topological for a valuation domain \(V\) of rank \(>1\) by describing first the polynomial closure of a pseudo-convergent sequence. Using this description they obtain a pseudo-convergent sequence \(E\) such that for some \(t\in E\) one has that \(\overline E\ne \overline{\{E\setminus t\}}\cup \overline{ \{t\}}\), implying that the polynomial closure is not topological. They also determine when two pseudo-convergent sequences have the same polynomial closure.