The model theory of separably tame valued fields (Q897758): Difference between revisions

In this paper, the authors establish several Ax-Kochen-Ershov (AKE) principles for the class of separably tame valued fields. Suppose \((K,v)\) is a Henselian valued field with residue field \(Kv\) and value group \(vK\). A finite extension \((E|K,v)\) of \((K,v)\) is called tame if {\parindent=0.7cm\begin{itemize}\item[1.] the ramification index \((vE: vK)\) is prime to \(p\), where \(p = \operatorname{char}Kv\) if \(\operatorname{char}Kv>0\), and \(p=1\) otherwise, \item[2.] the residue field extension \(Ev|Kv\) is separable, and \item[3.] the extension \((E|K,v)\) is defectless, i.e., \([E:K] = [Ev:Kv]\cdot (vE:vK)\). \end{itemize}} A separably tame field is a Henselian field for which all finite separable extensions are tame. The class of separably tame fields properly contains the class of separable-algebraically maximal Kaplansky fields, as well as the class of separably closed valued fields. The relevant languages for valued fields here are \(\mathcal{L}_{\text{VF}} = \{+,-,\cdot,{}^{-1},0,1,\mathcal{O}\}\), where \(\mathcal{O}\) is a binary relation symbol for valuation divisibility, and \(\mathcal{L}_{\text{Q}} = \mathcal{L}_{\text{VF}}\cup\{Q_m\}_{m=1}^{\infty}\), where the \(Q_m\) are the relations for \(p\)-independence, first introduced in [\textit{Yu. L. Ershov}, Sov. Math., Dokl. 8, 575--576 (1967; Zbl 0153.37201)] in the setting of separably closed fields. The main theorem of the paper is the following: Theorem. Let \((K,v), (L,v)\) be separably tame fields with the same characteristic \(p\) and finite \(p\)-degree. Then, \[ vK\equiv vL \;\& \;Kv\equiv Lv \;\Longrightarrow \;(K,v)\equiv (L,v) \] as \(\mathcal{L}_{\text{VF}}\)-structures, \[ (K,v)\subseteq (L,v) \;\& \;vK\prec vL \;\& \;Kv\prec Lv \;\Longrightarrow \;(K,v)\prec (L,v) \] as \(\mathcal{L}_{Q}\)-structures, and \[ vK\prec_{\exists}vF \;\& \;Kv\prec_{\exists} Fv \;\Longrightarrow \;(K,v)\prec_{\exists}(F,v) \] as \(\mathcal{L}_{Q}\)-structures where \((F,v)\) is an arbitrary extension of \((K,v)\). The proof involves establishing various embedding lemmas for valued fields, including at one point reducing to the case of tame valued fields and previous work by the first author [\textit{F.-V. Kuhlmann}, J. Reine Angew. Math. 719, 1--43 (2016; Zbl 1401.13011)]. As a corollary of the main theorem they derive several relative decidability results. They also give alternative proofs for well-known results on separable-algebraically closed valued fields and separable-algebraically maximal Kaplansky fields from [\textit{F. Delon}, Quelques propriétés des corps valués en théorie de modèles. (PhD thesis) (1982)]. The case of infinite \(p\)-degree is left open.