Linear topologies on a field and completions of valuation rings. (Q696208): Difference between revisions

Let \(A\) be an integral local ring and \(QA\) its quotient field. The linear topology on \(QA\) (\(A\)-topology for short) is given by the fundamental system of neighbourhoods of \(0\): \(\Sigma_A=\{a {\mathfrak M}(A) \mid a \in A, a\not= 0\}\). In the case where \(A\) is Noetherian, it is the usual \({\mathfrak M}(A)\)-adic topology. The results of this paper are new for the (interesting) case when \(A\) is not Noetherian. Let us denote \(\hat A:=\text{proj.lim} \;A/{\mathfrak a}\) \(({\mathfrak a} \in \Sigma_A)\). At the end of the paper, the author computes many meaningful examples of couples \((A,\hat A)\). My favourite one is the number 3 on page 70: \(A:=k[\mathbb Z^n]_{\mathfrak n}= k\oplus \bigoplus_{1 \leq i \leq n}t_ik(t_n,...,t_{i+1})[t_i]_{(t_i)}\), \({\mathfrak n}\) is the ideal of the Laurent series with constant coefficients equal to \(0\). \(\hat A = k \oplus t_1 k(t_n,...,t_2)[[t_1]]\oplus \bigoplus_{2 \leq i \leq n}t_ik(t_n,...,t_{i+1})[t_i]\), \(\hat A \not=k[[\mathbb Z^n]]\), rather unexpected. The main fundamental result is the equivalence: \(A\) is a valuation ring \(\Leftrightarrow \widehat A\) is a valuation ring. If it is so, the extension \(A \hookrightarrow \widehat A\) is immediate (same group of values, same residue field) and the morphism of ringed spaces \(\text{Zar}(Q\widehat A| A) \to \text{Zar}(QA| A)\) given by the restriction is a homeomorphism. \(\text{Zar}(QA| A)\) is the set of valuation rings of \(QA\) containing \(A\). The main idea of the proof is to look at \(A \hookrightarrow \hat A\) and to see when this map is continuous with respect to the \(A\) and \(\widehat A\)-topologies. In the case of the equivalence, it is continuous.