Products of finite abelian groups as profinite groups (Q2488319): Difference between revisions

A profinite group regarded as profinite structure is a pair \((G,\Aut^*(G))\), where \(G\) is a profinite group and \(\Aut^*(G)\), the structural group, is a closed subgroup of \(\Aut(G)\) respecting the inverse system; the structural group is standard if it consists of all such homomorphisms. The pair (and, by abuse of language, the profinite group) is called small if for all \(n<\omega\) there are only contably many orbits on \(n\)-tuples on \(G^n\) under the action of \(\Aut^*(G)\). These objects occur naturally in model theory as the quotient of a definable group in a small stable theory by its connected component. The author provides the first explicit examples: With the standard structural group, a direct product of finite groups, almost all of which are abelian of bounded exponent, with the inverse system given by the subgroups with trivial first \(n\) co-ordinates (for \(n<\omega\)), are small, and even \(m\)-normal and \(m\)-stable (see below). The same holds if almost all groups are \(p\)-groups and the standard structural group is replaced by a \(p\)-Sylow subgroup. The proof uses a variant of the quantifier elimination for modules. Profinite groups (and more generally profinite structures) were defined by \textit{L. Newelski} [J. Symb. Log. 66, 859--872 (2001; Zbl 0993.03049); Trans. Am. Math. Soc. 354, 925--943 (2002; Zbl 0985.03021)], who also introduced a notion of dimension for finite tuples (\(\bar a\) is independent of \(\bar b\) over \(\bar c\) if the orbit of \(\bar a\) over \(\bar b\bar c\) is open in the orbit of \(\bar a\) over \(\bar c\)) and proved that it had properties analogous to stability-theoretic independence. Theories where dependence is well-founded are called \(m\)-stable (similar to superstability); \(m\)-normality is an analogue for one-basedness (the canonical base of a type/orbit is algebraic over any one realization). He conjectured that a small profinite group should have an open abelian subgroup; this was proved by \textit{F. O. Wagner} [Fundam. Math. 176, 181--191 (2003; Zbl 1025.03027)] in the \(m\)-stable case; the general case is still open.