Counterexamples to a conjecture on relative categoricity (Q750425): Difference between revisions

The paper is a brief supplement to the paper of \textit{W. Hodges, I. M. Hodkinson} and \textit{D. Macpherson} [Ann. Pure Appl. Logic 46, No.2, 169- 199 (1990; Zbl 0699.03016)]; its main aim is to refute a conjecture made at the end of that paper. The presented result is based on the following theorem proved in the present paper. Theorem: There exists a profinite group G (in fact, a Cartesian product of countably many finite groups) with a non-trivial finite normal subgroup F which has a complement in G, and such that any complement to F in G is dense in G (and therefore not closed in G). Moreover, the methodology used by the authors allows to construct a structure B with the following properties: (a) There exists a countable \(\omega\)-categorical structure C such that the automorphism groups of B and C are isomorphic as groups but not as topological groups (so and C are not bi-interpretable); (b) in the terminology of \textit{M. Rubin} [On reconstruction of \(\omega\)-categorical structures from their automorphism groups, Preprint (1988)], the structure B has no weak \(\forall \exists\)- interpretation. The proper existence of the structure B answers another open question presented by Rubin [loc. cit.].