Normal hyperimaginaries (Q403407): Difference between revisions

A theorem of \textit{D. Lascar} and \textit{A. Pillay} [J. Symb. Log. 66, No. 1, 127--143 (2001; Zbl 1002.03027)] states that in a simple theory, any bounded hyperimaginary (i.e., class of a countable tuple modulo a type-definable equivalence relation whose orbit under the automorphism group is bounded) is interdefinable with a sequence of finitary hyperimaginaries (i.e., such classes of finite tuples). Their proof uses the Peter Weyl theorem which states that a compact Hausdorff group is an inverse limit of compact Lie groups. In this paper, a more model-theoretic proof is given, which nevertheless still uses the Peter-Weyl theorem. The authors introduce the notion of a \textit{normal} hyperimaginary as one whose stabilizer is a normal subgroup of the full automorphism group. Equivalently, a hyperimaginary is normal iff it is interdefinable with all its model-theoretic conjugates. This in particular implies that it is bounded. It is easy to see that every normal hyperimaginary is interdefinable with a sequence of finitary hyperimaginaries; the authors prove the theorem of Lascar and Pillay [loc. cit.] under the hypothesis that \((\pmb\emptyset)\) is interdefinable with a sequence of normal hyerimaginaries. This hypothesis is a consequence of the Peter-Weyl theorem. In a second section, the authors show moreover that every closed subset of a Kim-Pillay space has a hyperimaginary canonical parameter.