CONSTRUCTING AN ALMOST HYPERDEFINABLE GROUP (Q4658682)

The paper extends the group configuration theorem from stable to simple theories. In the stable setting, the theorem basically says that, whenever a dependence/independence situation arises, then there is some non-trivial group behind it. A possible proof strategy within stable theories consists of two key steps: first obtain a generic group chunk whose elements are germs of generic functions and whose product is the composition, and then apply the Weil-Hrushovski generic group chunk theorem. In the simple setting, the third author suitably extended the latter theorem in his paper [``Hyperdefinable groups in simple theories'', J. Math. Log. 1, 125--172 (2001; Zbl 0980.03040)], while the first author dealt with the former step in his paper [``Group configurations and germs in simple theories'', J. Symb. Log. 67, 1581--1600 (2002; Zbl 1043.03029)]; he obtained there a generic {polygroup} chunk, meaning that the generic product is defined up to a bounded non-zero number of possibilities. So a strategy to accomplish the proof and find a generic group chunk may just start from here: divide the polygroup chunk by some suitable equivalence relation and then give a blow-up argument to extend the elements properly. By the way, a dividing equivalence relation as just described is called {core}. The basic obstacle the authors have to overcome relies on the fact that the core equivalence relations they handle are not type-definable. Anyway, it is observed that they satisfy several nice properties, and indeed the authors call them {almost type-definable} just to emphasize this behaviour. More generally, the notions of {almost hyperdefinable polystructure} and {almost hyperimaginary} are introduced and a reasonable model theory is developed about them, including an existence theorem of generic elements for almost hyperdefinable polygroups in simple theories. At this point blow-up arguments are used to transform a coreless almost hyperdefinable polygroup chunk into an almost hyperdefinable group chunk. Then the final step of the proof is to show that the group chunk theorem preserves almost hyperdefinability. The paper includes a clear introduction, illustrating in detail the problem, the genesis of the proof and the hurdles it required to pass.