Superstable subgroups of \(\text{SL}_2(K)\) and \(\text{PSL}_2(K)\). (Sous-groupes superstables de \(\text{SL}_2(K)\) et de \(\text{PSL}_2(K)\)) (Q2491806)

The first author has shown [``Sous-groupes de SL\(_2(K)\) sur un corps superstable'', in: L. Bélair (ed.) et al., Model theory and applications. Quaderni di Matematica 11, 297--306 (2002; Zbl 1084.03033)] that a definable subgroup of \(\text{SL}_2(K)\), where \(K\) is a superstable field, is either soluble or a conjugate of SL\(_2(k)\) for some definable subfield \(k\); this generalized a result by the second author [``Quelques modestes remarques a propos d'une conséquence inattendue d'un résultat supprenant de Monsieur Frank Olaf Wagner'', J. Symb. Log. 66, 1637--1646 (2001; Zbl 1005.03039)], where \(K\) was assumed to be of finite Morley rank. In this paper the result is generalized to groups superstable \textit{in the group language}, i.e.\ the underlying field and the linear action need not be definable, with the proviso that the conjugation may take place in \(K(\alpha)\), where \(\alpha^2\in K\). The main new ingredient is the definition of a generic copy of the small field \(k\) on the set of traces of \(G\) via the trace formula established by the second author [loc.cit.]. Consequently, the methods seem restricted to dimension \(2\). With slightly more effort, the analogous result is shown for a superstable subgroup of PSL\(_2(K)\). (This would be immediate if \(K\) were known to be superstable, but for \(G\leq \text{SL}_2(K)\), superstability of \(G/Z(\text{SL}_2(K))\) need not a priori imply superstability of \(G\).) As corollaries, the authors obtain that a superstable subgroup of a free group is commutative, and that there is no linear bad group of rank \(3\).