Locally finite weakly minimal theories (Q1182488)

The aim of the paper is the consideration of weakly minimal sets \(M\) of nonalgebraic strong type \(p\) satisfying the following condition: \((*)\) For any \(N\geq M\) and finite \(F\subseteq p(N)\) there is a finite \(G\subseteq p(N)\) such that \(\text{acl}(F)\cap p(N)=\bigcup\{\text{acl}\{g\}: g\in G\}\cap p(N)\). That means, the geometry on \(p\), which arises from the algebraic closure operation acl, is locally finite. The presented investigations are related to papers of Andler, Buechler, Hrushovski and Shelah. After the definition of the basic notions and fixation of the notation a description of affine and projective spaces over finite fields is given: The geometry \((M,\text{acl})\), where \(M\) is an \(\aleph_ 0\)-categorical strictly minimal structure, is isomorphic either to the degenerate geometry on \(M\), to an affine geometry of infinite dimension over a finite field \(F(n)\), or to a projective geometry of infinite dimension over \(F(n)\). An affine (projective) structure over \(F(n)\) is defined as a structure \(M\) with a ternary predicate \(R\) and a 4-ary predicate \(Q\) (ternary predicate \(S\)) satisfying a number of axioms. Let \(M\) be a weakly minimal structure and let \(p\) be a strong type satisfying \((*)\) and having nontrivial dependency relation. Then there is an almost 0-definable equivalence relation \(\theta\) on \(M\) with finite classes and \(\psi(x)\in p\) such that \(\psi(M)/\theta\) is an affine or projective structure over a finite field. If \(M\) is moreover unidimensional, then a global structure theorem for \(M/\theta\) can be proved in several steps. The affine case and the projective case are considered separately. Via some partial results one obtains the main Theorem: ``Suppose \(M\) is a weakly minimal structure and \(p\) is a strong type satisfying \((*)\) of projective (affine) character. Then there is \(\theta\) an almost 0- definable equivalence relation with finite classes, and \(\sigma\in p\) such that \(\sigma(M)/\theta\) is a projective (affine) space over a finite field. There is some collection of algebraic points of \(\sigma(M)/\theta\) (and possibly \((\sigma(M)/\theta)^ 2/Q)\) and some collection of almost 0-definable subspaces of \(\sigma(M)/\theta\), and \(\sigma(M)/\theta\) has essentially no other structure.'' Finally the author considers the case where the structure \(M\) is in addition unidimensional. This implies that \((*)\) holds on every strong type.